Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
© 2015 Jacob Lee Meyer
DEVELOPMENT OF AROMATIC POLYESTERS FOR HIGH
PERFORMANCE APPLICATIONS AND USE OF INTERCHAIN
TRANSESTERIFICATION REACTIONS AS A SOLID-STATE
FABRICATION TOOL
BY
JACOB LEE MEYER
THESIS
Submitted in part ial fulfillment of the requirements
for the degree of Master of Science in Materials Science and Engineering
in the Graduate College of the
University of Illino is at Urbana-Champaign, 2015
Urbana, Illinois
Adviser:
Professor James Economy
ii
ABSTRACT
Deposit ion and cure of aromatic thermosetting copolyester (ATSP) oligomers is
explored across several potent ial schemes: plasma spray, solvent -borne, and as an
electrostatic powder coating. It is generally found that all methods are potent ially
viable as low frict ion, low wear coat ings but that ATSP as an electrostatic powder
coating is the simplest and highest performing deposit ion mechanism in terms of
wear rate and provides the highest glass transit ion temperatures in an
electrostatically-deposited polymeric powder coating known in literature or
commercially. Low production of wear debris coupled with stable and high
mechanical property films transferred to the tribopair counterface (as seen via
energy-dispersive X-ray spectroscopy) lends evidence that during tribological
experiments that produce contact temperatures of up to 320°C that interchain
transesterificat ions (ITR - a class of solid state chemical react ions available to
aromatic polyesters) are responsible in part for low wear rates via ATSP
reincorporating its own wear debris and development of a stable transfer film.
Addit ionally, it is found that a mixed liquid crystalline/amorphous oligomer set will
phase segregate and produce surface texture corresponding to the phase segregated
ordered region. This surface texture and the high glass transit ion temperature of
ATSP enables the in-situ format ion of micro-reservoirs of lubricant during
lubricated tribotesting.
Micromechanical experiments were used to assess the contribut ion of cohesive wear
to ATSP resins as well as several neat polymer and commercially deposited
iii
coatings. In instrumented scratch experiments ramped up to 80mN in applied force
with a 4.3 μm conispherical indenter t ip, it is found that ATSP ex hibits a uniquely
high tendency towards what is termed “elast ic recovery” wherein the scratched
region of the coat ing surface recovers most of the deformed depth during retrace
experiments. As well, the wear mode of amorphous ATSP coat ings maintains this
performance at applied loads significant ly in excess of any other observed coat ing.
It is hypothesized that the crosslinked structure and potential for the aromatic
polyester chains to deform via crankshaft motion about the ester bond may be
responsible for this interest ing mechanical feature.
Vacuum assisted resin transfer molding is demonstrated as a potential fabricat ion
technique for cont inuous fiber ATSP composites. The potent ial for ATSP to be the
first weldable thermoset material is also explored. Fully cured laminae of
ATSP/carbon fiber are fused into a single workpiece via solid -state interchain
transesterificat ion react ions. Resin-dominated mechanical properties of the
composite such as interlaminar shear strength and the Mode I fracture toughness of
the composite at the bondline adhered via ITR. Results indicate interlaminar
properties beyond that of tradit ional epoxy and polyimide resins. Addit ionally,
cryogenic thermal cycling experiments are used to assess the composite’s resistance
to microcracking which indicates excellent resistance to microcracking as well as
potent ial repair schemes for ATSP composites. Single fiber fragmentation testing
was used to determine the interfacial shear strength of the fiber/matrix bond. The
high strength here and observat ion via scanning electron microscopy of all fibers
iv
ent irely coated in resin suggest that cracking in ATSP composites predominant ly
occurs within the resin phase. The lack of low energy surfaces at fiber/matrix
interface to propagate suggests a rat ionale to the high Mode I fracture toughness
observed for ATSP composites.
Mechanical and thermal performance of ATSP foams and fully dense structures
recycled from the foams via interchain transesterificat ion react ions are described.
Tribological results of the fully dense structures when filled with PTFE lubricat ing
addit ives suggest applicat ion as a low-frict ion and low wear material. Several novel
addit ive fabricat ion schemes and a process to weld aromatic polyester composites
are also suggested as future areas of research.
v
To My Wife Amanda and My Son Oliver
vi
TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION & BACKGROUND……………………………1
CHAPTER 2: OLIGOMER SYNTHESIS…………………………………………20
CHAPTER 3: DEVELOPMENT OF PRACTICAL COATING
SCHEMES UTILIZING ATSP…………………………………………………….33
CHAPTER 4: MICROMECHANICAL PROPERTIES AND ELASTIC
RECOVERY OF AROMATIC THERMOSETTING COPOLYESTERS………100
CHAPTER 5: ADVANCED CHARACTERIZATION OF CONTINUOUS
FIBER / ATSP COMPOSITES……………………………………………..………123
CHAPTER 6: INTERCHAIN TRANSESTERIFICATION AS A SOLID-
STATE FABRICATION TOOL……………………………..………………..……170
REFERENCES……………………………….………………………………………207
1
CHAPTER 1
INTRODUCTION & BACKGROUND
1.1 Background
Aromatic polyesters, polyamides, and polyimides are established classes of polymers generally
incorporating an aromatic structure as the basis of their mesogenic units. These semi-rigid
structures have seen a broad range of commercialization, with their greatest commercial success
as polymeric fibers such as most notably the lyotropic LCP poly(p-phenylene terephthalamide)
(PPTA), which is marketed as Kevlar and Twaron [1,2]. The thermotropic copolymer of 4-
hydroxybenzoic acid (HBA) and 2,6-hydroxynaphthoic acid (HNA) is also marketed under the
trade name Vectran [1,2]. Aromatic polyesters are sold for molding purposes notably include the
copolymer of terephthalic acid, biphenol, and p-hydroxybenzoic acid known as Xydar and
Vectran also marketed as Vectra[1]. The related homopolymer of HBA, commercially known as
Ekonol, was the first to enter the commercial sphere after the original discovery and
development work at Carborundum Corporation by Economy and co-workers of the
solution/suspension phenyl ester route [3].
The high mechanical properties of these materials are seen to arise from their orientational self-
ordering [1,38] and high rigidity of their constituent units. Examining one of the prototypical
and most common constituent of aromatic polyesters, poly(HBA) as shown in Figure 1, the
phenylene units are prevented from rotating and conformational freedom is only available
through crankshaft motion about the ester bond [38].
2
In addition to their robust mechanical properties [1-5], aromatic polyesters are also known to
have excellent chemical resistance [4], high heat deflection temperatures and flame
resistance[9,42], and hydrolytic stability [7]. These features make them attractive as matrices in
high performance composites. To date though, no all-aromatic polyester is known to have been
successfully marketed as a matrix for continuous fiber reinforced composites, in spite of the lack
of a generally accepted polymeric matrix material for service temperatures above 135°C [41].
1.2 Aromatic Polyesters
Aromatic polyesters represent some of the earliest reports on polymeric materials,
with Klepl reporting on residue he called p-oxybenzid during his synthesis of the
dimer and trimer of HBA in 1883 [43]. Fischer later confirmed the synthesis of the
dimer and tr imer in 1909 [44]. Work on aromatic polyesters laid dormant unt il
Gilkey and Caldwell reported on the polymerizat ion of p -acetoxybenzoic acid and
m-acetoxybenzoic acid [45]. They erroneously concluded the polymer of HBA as
unstable due to their melt polymerizat ion producing the phenyl ester of p-
Fig. 1.1: Schematic of poly(HBA) repeat unit with bond angles and lengths expressed.
From [37].
3
hydroxybenzoic acid by an acid catalyzed condensat ion of the phenolic hydroxyls or
acetoxy units [46]. Economy first successfully produced the homopolymer of HBA
(PHBA) in 1963 by use of a phenyl ester to cap the carboxylic acid and use of a
heat exchange medium to control the react ion temperature [ 3]. The highly
crystalline product was thermally stable in air at 350°C and was by 1970
commercialized for use in thermal spray and high energy rate forging [ 2].
This was soon followed by the development of the copolymer of PHBA and
biphenol terephthalate (BPT) which by 1971 was commercialized as Ekkcel [1]. It
was observed early in the work that the 2:1 ratio of PHBA:BPT yielded an inject ion
moldable grade that melted at 408°C while the 1:2 rat io copolymer yielded a
compression moldable grade.
The 3:2 copolyester of p-hydroxybenzoate and ethylene terephthalate was
commercialized by Eastman-Kodak in the mid-1970s, though it was withdrawn soon
thereafter due to poor thermal stability. The 3:7 copolyester of HNA and PHBA,
sold as Vectra and Vectran as ment ioned above, was commercialized in 1980 by
Celanese. This copolymer has seen ut ilization in both moldable products and as
fibers, however its low glass transit ion temperatur e of 150°C and degradat ion
during annealing [51] has limited its further deployment.
4
Fig. 1.2. Structures of previously commercialized aromatic polyesters.
1.3 Aromatic Thermosetting Polyester (ATSP)
Aromatic thermosetting polyesters (ATSPs) are a relat ively new family of
thermosetting polyesters made from low cost monomeric precursors. Commercially
available monomers (HBA, trimesic acid, hydroquinone, and isophthalic acid) with
phenyl hydroxyl-capped monomers acetylated to acetoxy form via acet ic anhydride
with sulfuric acid catalyst. The acetoxy and carboxylic acid capped aromatic
monomers undergo melt condensat ion reactions at 100-270°C, depending on the
molecular weight of the desired oligomeric system, to yield either aceto xy-
5
terminated or carboxylic acid-terminated melts [4]. The oligomers are blended then
cured at 250-330°C to yield ATSP (Figure 2).
A number of oligomeric systems with varying funct ionality and molecular weight
have been developed and characterized. The oligomers can be tailored to be liquid
crystalline by adjust ing the molar rat io, thus providing unique advantages compared
to convent ional polymers. In addit ion, the number of funct ional groups can be
modified to affect the crosslink density. Varying the molecular weight adjusts the
mechanical properties of the neat resin and composite. Table 1 shows some of the
typical oligomeric compounds that have been synthesized along with their softening
points.
R1
O
OH+
CH3
O
OR2 R1
O
OR2 +
CH3
O
OH
Acetoxy
end group
Carboxy lic
acid end
group
Ester backbone
of ATSP
Reaction
byproduct
OO
HOO
O
O
OO
OHO
O
OO
OO
O
O
O
O
OH
OO
O
OO
HO
OO
O
O
n
O
O
OO
O
O
OO
O
OO
OO
O
O
OO
OO
O
O
OO
O
O
O
O
O
O
O n
(a)
(b)
Carboxylic Acid End
Group (C1)
Acetoxy End
Group (A1)
By-productEster Backbone
ATSP
Carboxylic Acid End
Group
Acetoxy End
Group
Trimesic acid
(TMA)
4-acetoxy-benzoic acid
(ABA)
Isophthalic acid
(IPA)
Hydroquinone diacetate
(HQDA)
3
Fig. 1.3: Proposed structure of ATSP. From [51]
6
Table 1.1: Properties of ATSP oligomers. From [4].
Oligomers marked with * exhibit liquid crystallinity in melt
ATSP can be produced in either amorphous or liquid crystalline formulat ions by
select ion of appropriate monomer feed ratio to select crosslink density. Oligomers
which do not display birefringence individually in their melt form amorphous
polymers when cured while those that do display birefringence produce polymers
which are birefringent after full curing [7,8]. Birefringent oligomers have similar
molecular weights as isotropic oligomers but with a lower feed rat io of the
crosslinker (TMA) which results in a longer chain length between crosslink s, as
shown in Table 2.2. Figure 1.4 displays the difference in birefringence between
amorphous and liquid crystalline ful ly cured ATSP.
7
1.4 ATSP Physical Properties
As described above, ATSP is a high temperature resin that has unique properties [ 1-
5,7,9]. Prior experiments demonstrate ATSP has better thermal stability and low
moisture pickup compared to epoxies, with properties equivalent to those of best
performing thermoset polyimides [12-15] and can be repaired more easily than
either [21]. ATSP is stable in air at 350 C and in nitrogen at 425 C [4,20] whereas
most thermally stable epoxies decompose at 170-190 C in air or nitrogen. The
moisture pickup of the resin is relat ively low (0.3 wt%, as compared to 2.3% for
epoxy and 2.6% for polyimides) [12-15], increasing the potent ial durabilit y against
physical ageing. ATSP also shows outstanding f lame [42] and ablat ive character [9]
which may be of ut ility in the design of high performance aerospace structures. In
these experiments, ATSP composites formed a stable char and evidenced no
Fig. 1.4: Observation of birefr ingence of ATSP fiber/resin compositions via cross -
polarized light microscopy. From [51].
8
interlaminar delaminat ion after exposure to oxyacetylene torch events while Parkar
et al [9] demonstrated its potential suitability as a heat shield material for both
Earth and Mars re-entry via COIL-type laser ablat ion experiments at marker flux
rates.
The liquid crystallinity of the matrix [36] and the tendency of the aromatic
polyester repeat unit to be oriented [37-40, 49] is hypothesized to aid in reducing
thermally induced interfacial stresses when used as matrix for carbon fiber
composites [12,18,36,50-51] via local matching of the coefficient of thermal
expansion around the carbon fiber . It has been observed that birefringence (and
thereby implying orientat ional order) in the resin is preserved in regions
immediately surrounding carbon fibers [17,50] even in temperature regimes where
the resin would normally be in an amorpous state. An example of this is illustrated
in Fig. 1.4 wherein a normally amorphous C1A1 resin shows induced birefringence
surrounding a carbon fiber. Further, the order parameter in LCPs surrounding
carbon fibers are enhanced to approximately that of the substrate carbon fiber [55]
and was found to be independent of fabricat ion approaches . Preliminar y
experiments [51] for coefficient of thermal expansion in ATSP/C composites
demonstrated a reduced transverse coefficient of thermal expansion a nd a
significant ly reduced difference between transverse and longitudinal as compared to
epoxy/C composites of comparable resin CTE, ident ical fiber and weave type, and
volume fract ion of resin. Such an effect could be ant icipated to be of utility in
9
cryogenic or high temperature applicat ions where thermal cycling was encountered
such as in reusable composite cryogenic fuel storage tanks .
Table 1.2. Summarized ATSP (C1/A1) Physical Properties [7,10,11,42,47]
1.5 Interchain Transesterification Reactions
One of the most salient feature of this system is an ability to form adhesive bonds
even after being cured. The ester crosslinks impart to the resinthe ability to undergo
interchain transestericat ion react ions (ITR) [14]. Interchain transestericat ion
react ions (ITR) are unique to polymers with ester linkages and have been found to
aid in the format ion of bonds between molecular chains [5, 10, 21]. This can be
exploited to obtain adhesive bonds between polymers with ester linkages . Surface
react ions can occur at high temperatures to form strong chemical bonds.
10
Interchain trans react ions, of which ITR is a const ituent member, have several
dist inct ive features. First ly is that the functional groups present before and after the
react ion are ident ical [56]. Trans react ions (also known as chemical interchange
react ions or exchange react ions [96]) are not unique to polyesters and are, in fact,
conceptually plausible in any condensat ion polymer, though only a few have been
ident ified. Other interchange react ions so far ident ified include those those among
polyamides between amide and amine functional groups and between amide groups
[57]. Interchange react ions have been ident ified among siloxanes in the presence of
strong base or sulfuric acid [57-58]. Interchange react ions in polysulfides are
known to occur via 1) between thiol and disulfide bonds, 2) adjacent disulfides, and
3) between inorganic and organic disulfides, which produces mercaptan [57,59].
These have been proposed as a stress-relief mechanism in vulcanized rubbers under
strain at elevated temperatures [60].
Exchange react ions within the polyester backbone can be classed into three types:
The first type is an exchange react ion between a hydroxyl end group with a
neighboring ester that is termed an alcoholysis react ion [56] and is completely
reversible at every step. This proceeds by nucleophilic addit ion of the hydroxyl to a
carbonyl group of an ester linkage. This induces a posit ive charge to the hydroxyl
oxygen and a negat ive charge to the carbonyl oxygen. The intermediate structure
expels an alchohol via electron transfer from the negat ively-charged oxygen which
captures a proton from the original hydroxyl. The final result is that exchange
occurs at the carbon-oxygen single bond in the original ester [62]. Flory previously
11
demonstrated that presence of an acid catalyst enabled ready react ion of copolymers
in excess decamethyle glycol and adipic acid, which produced hydroxyl terminated
polymers. This was determined by inferrin g a decrease in molecular weight to the
copolymer by observing a viscosity decrease with decamethylene glycol monomer
addit ion to the high MW polymer melt. This type exchange react ion in polyesters is
commercially exploited for the synthesis of polyethylen e terephthalate from
ethylene glycol and terephthalic acid[61].
Figure 1.5 and the lower diagram in Figure 1.6 demonstrates the second type of
exchange react ion observed for polyesters: acidolysis. This react ions proceeds by
nucleophic addit ion in a similar manner described for alcoholysis above wherein an
acyl oxygen on a carboxylic acid adds to a carbonyl of a backbone ester. Again, the
tetrahedral intermediate of the original polyester acyl oxygen eliminates an ester
linkage formed between the original ester oxygen and the carbonyl of the original
carboxylic acid. This also produces a new carboxylic acid [62]. An example of use
this class of react ion is the react ion of PET with with 4-acetoxybenzoic acid (4-
ABA) to produce a copolymer [64]. It was assume d in this study that only
acidolysis proceeded, however that may not be a valid assumption due to the
possibility that the third class of ester exchange react ions occurred between the
acetoxy unit of the 4-ABA and the ester backbone units of the high MW PE T.
Ester-ester exchange, known as esterolysis, is the third mechanism for ITR in
polyesters. An example react ion for aromatic polyesters is show in the upper
portion of Figure 1.6. Rather than the nucleophilic pathways for alcoholysis and
12
acidolysis, this exchange react ion is thought to proceed by what is termed an
associat ive react ion mechanism [66-67]. Two carbonyls of adjacent ester bonds
experience format ion of an associat ion complex and the acyl groups of the original
esters are switched. While Flory and others [57,68,69] have suggested that evidence
indicates that esterolysis proceeds very slowly, contradictory evidence exists
[67,70,71] for exchange of PET and bisphenol-A diacetates. These were determined
both through NMR and viscosity changes.
Although three react ions have been shown to exist, obviously they cannot all have
the same rate constant. Complicat ions in determining the rate constant for each
mechanism arise from difficult ies in separating the rates of ITR from condensat ion
esterificat ions as well as isolat ing the individual react ion contribut ions
(alcoholysis, acidolysis, and esterolysis). Flory inferred that rates of alcoholysis
and esterificat ion were approximately the same for similar condit ions and without
the addit ion of an external catalyst for polymers of adipic acid and decamethylene
glycol. Esterificat ion rates were determined by mixing equimolar quant it ies o f
decamethylene glycol and adipic acid and monitoring the change in viscosity.
Alcoholysis rates meanwhile were determined by measuring the viscosity change
upon addit ion of decamethylene glycol to decamethylene adipate polymers. He
found that acidolysis rates were approximately an order lower than for alcoholysis
[68]. Hamb [70] and Tijsma [67,71] however est imated that esterolysis was
significant ly faster than acidolysis and of the same order as alcoholysis.
13
Later technological advances aided in the determinat ion of transesterificat ion rates
within a polymer. Kugler [73] determined that alcoholysis and esterolysis were
responsible for decrease in length of deuterated and protonated segments via small
angle neutron scatter (SANS) with no associated change in viscosity which
indicates a randomizat ion process. These experiments further suggest this follows
an Arrhenius process as the decrease in the segment length was as a funct ion of
both temperature and t ime [73]. Arrighi observed a similar result via SANS study of
dimethylbenzalazine copolyesters [74] and suggested that ITR funct ioned to
produce the most probable MW distribut ion of the polymer (broadening
polydispersity index, PDI) while leaving the number average molecular weight M N
unchanged. This randomizat ion is a natural consequence of the property of ITR to
leave the count of funct ional groups unchanged. MacDonald et al [113] used SANS
against a part ially deuterated wholly aromatic polyester of high molecular weight to
determine an act ivat ion energy of of 157 kJ mol-1
.
Economy and Schneggenburger [53,54,75] demonstrated via 13
C NMR that a
mixture of 4-hydroxybenzoic (HBA) and 2,6-hydroxynapthoic acid (HNA) powders
homopolymers would undergo extremely rapid randomizat ion via ITR (within 10
seconds at 450°C) to produce a HBA/HNA copolymer. This had drast ic physical
effects on the sample: the melt ing point of the copolymer was d rast ically lower than
either of the homopolymers. The HNA/HBA diad sequences were obvious via the
13C NMR spectra. From the NMR data and physical observat ions a react ion rate of
10000 transesterificat ions per molecule per minute – a tremendously fast rate.
14
These observat ions enabled subsequent studies to demonstrate the ability to alter
average molecular weight and distribut ion via blends of high MW polymers [76 -77]
and showed the potentional for localized post -processing of polyesters [78-79].
Frich and Economy demonstrated that a crosslinked and fully cured network
polymer (ATSP) coated onto an aluminum alloy substrate and then bonded at 330°C
to a counterpart coated substrate demonstrated excellent cohesion. A lap shear
strength of up to 20 MPa was demonstrated, comparable to commercially available
epoxy adhesives of similar of bondline thickness. Furthermore, it was observed that
the failure mechanism of the lap shear coupons was ent irely adhesive at the
aluminum/ATSP bondline rather than at the ATSP/ATSP bondline. This suggested
high strength for this joining mechanism. An associated neutron reflect ivity study
by Frich and Economy [10] an interdiffusion distance of 30 nm across a bondline of
two ATSP thin films adhered via ITR in a similar manner as above, an order lower
than that predicted by Kramer as necessary for adhesion in the case of strict ly
physical diffusion of rigid chains across an interface [ 143]. As these were
previously fully cured films, a strict ly physical diffusion was not a viable
mechanism. Frich and Economy also demonstrated via 1H NMR that deuterated
oligomers of ATSP undergo exchange with protonated counterpart oligomers [48].
This body of work demonstrated that the ITR mechanism was a chemical process
and that it can be used to fuse two adjacent surfaces at the bondline in a solid-state.
This suggested of ut ility for manufacturing, bonding, and repair of thermosett ing
polyester composites (ATSP/C). Lopez et al. have shown that cured lamina can be
15
bonded together using ITR [5]. Thick consolidated sect ions of ATSP are obtained
by bonding single laminae under heat and pressure. It was also demonstrated that
ATSP films could be bonded in the solid-state to polyimide films of commercia l
Kapton [102,103]. This suggests an enabling fabricat ion technique for
microelectronic and micromechanical devices which may require an all -solid state
process. Interest ing as well is that the experiments indicated a diffuse boundary
layer between the Kapton and ATSP films [103]. A precise ident ificat ion of an
exchange process versus a physical process was not produced in that study, but the
potent ial of an imide-ester exchange react ion remains a viable hypothesis from that
experiment. This also opens the possibility that an aromatic imide -ester copolymer
as in Xu [106] could undergo ITR.
Huang and Economy use ITR as a solid-state bonding process for thin films
deposited on copper foils [104]. Huang [107] and Zhang [19,20] demonstrated that a
fine powder of ATSP could be consolidated via ITR at 330°C and under applied
pressure and that this evidenced a relat ively high glass transit ion temperature as
well as excellent tribological propert ies when a high concentrat ion of PTFE [19,20]
was added as a lubricat ing filler though a relat ively poor oxidat ive stability p rofile
was obtained due to limitat ions of the ATSP powder synthet ic route chosen in that
study. Recent studies by Calleja [96] and Capelot [112] used ester units bound
within monomeric species that were cured using an epoxide funct ional group.
Exchange react ions were then used to produce a chemically cont iguous networks
that could be dynamically deformed. These studies as well demonstrated that
16
catalysts used in commercial small molecule syntheses employing
transesterificat ions were pert inent to solid-state ITR in that they increased the
observed rate of exchange react ions.
Fig. 1.5. Example chemical react ion at surface between carboxylic acid and ester
unit via acidolysis. An example react ion for surface-surface bonding of crosslinked
aromatic polyesters.
17
Fig. 1.6. Proposed interchain transesesterificat ion react ion mechanisms in aromatic
polyesters for esterolysis and acidolysis.
18
Fig. 1.7. Overall process schematic for fabricat ion of ATSP structures. Blue
indicates input materials, red indicates output materials.
With these available mechanisms and several pathways that will be demonstrated in
this document, we put forth the following overall process schema for fabricat ion of
ATSP structures (Figure 1.7). A mechanist ic diagram for ITR is show in Figure 1.6.
Chapter 2 will discuss oligomer synthesis and network parameters of several cured
structures. Chapter 3 will describe the coating deposit ion methods and macro -
tribological characterizat ion of ATSP coatings and covers evidence for a self -
healing effect for ATSP coatings under aggressive and high temperature tribological
19
coatings via the ITR mechanism. Chapter 4 will cover micromechanical
characterizat ion of ATSP coat ings and elucidates at the micro-scale some observed
features of ATSP . Chapter 5 reports on progress towards out-of-autoclave synthesis
of cont inuous fiber ATSP composites and advanced characterizat ion of laminates
produced via ITR. Chapter 6 will discuss high specific strength and st iffness ATSP
foams and their recyclability into fully dense art icles with high mechanical and
tribological performance and aw well proposes several fabricat ion techniques using
ITR.
20
CHAPTER 2
OLIGOMER SYNTHESIS
2.1: Oligomer Synthesis
Oligomer synthesis followed a process derived from Economy and Frich [47] with
some modificat ions as described for new ATSP oligomers below.
Monomers were acquired from Alfa-Aesar at 99% grades except ing: biphenol,
which was acquired from Akron Polymer Systems at 99% grade; perfluorosebacic
acid, which was acquired from ExFluor Research Corporation at 95% grade; and
acet ic anhydride, which was acquired from Sigma-Aldrich at 99.5% grade.
To increase the glass transit ion temperature of the resin, the hydroquinone, which is
the ATSP component most thermally sensitive [2,114,115] (primarily by react ions
yielding phenoxy species and ketene) , was replaced by biphenyl units that are more
thermally stable (see Figure 2.1) and have lower vapor pressures at el evated
temperature thereby reducing losses from ITR at exposed surfaces. Hydroquinone
diacetate was replaced with biphenol diacetate in carboxylic acid - and acetoxy-
capped oligomers to increase the oxidat ive stability of oligomers. Acetoxybenzoic
acid (ABA) and biphenol diacetate (BPDA) were prepared by acetylat ion of p -
hydroxybenzoic acid and dihydroxybiphenyl, respect ively. The designated CB
oligomers were synthesized by melt -condensat ion of TMA, ABA, IPA and BPDA
(molar rat io 1:3:2:2 respect ively) at 260°C in the reactor. AB oligomers were
21
synthesized similarly with TMA, ABA and BPDA taken in the molar rat io 1:3:3. CB
and AB oligomers were then mixed and cured to make CBAB cured powders.
Fig. 2.1. ATSP backbone locat ions off hydroquinone monomer most sensit ive to
side react ions.
Biphenol diacetate (hereinafter BPDA) was synthesized by acetylat ion of bipheno l
(hereinafter BP) analogously to the synthesis of 4 -acetoxybenzoic acid. In this case,
500 g of BP was mechanically st irred in 850 mL of acet ic anhydride (molar rat io of
about 2.7:8.9) in a cylindrical vessel in an ice-water bath at 10°C at which point 2-3
drops of sulfuric acid was added to catalyze the acetylat ion react ion. The solut ion
temperature immediately increased to 80-85°C due to the exothermic react ion. After
allowing the solut ion to cool to room temperature, BPDA was precipitated out with
dist illed water. BPDA was then filtered, washed with copious volumes of dist illed
water and dried in a convect ion oven at 70 °C for 48 hours. The react ion yield was
above 95% as determined by integrat ion of 1H (proton) nuclear magnet ic resonance
(NMR) spectra following methodology from Guo [99] .
22
To produce carboxylic acid end-capped oligomer CB, 129.4 g TMA, 333 g BPDA,
204.7 g IPA, and 332.9 g ABA were mixed in a 2 L cylindrical reactor flask. The
flask was equipped with a three-neck head connected to an inlet inert gas, a screw-
type impeller driven by an overhead mechanical st irrer, a J -type thermocouple, and
an Allihn-type condenser with a 3-way valve to offer a toggle between reflux and
dist illat ion modes. The reactor was cont inuously purged with argon while emplaced
within a Glas-Col aluminum-housed electric heat ing mant le operated via
temperature controller with the thermocouple operat ing as feedback. The reactor
was heated to260 °C for 30 min to obtain a low-viscosity melt during which st irring
was maintained at 300 revolut ions per minute (rpm) and which evidenced a
substant ial acet ic acid by-product which was refluxed during this stage. After
refluxing for 30 minutes, the condenser was toggle to dist illat ion mode and acet ic
acid condensat ion by-product was collected in an Erlenmeyer flask and the mass of
the acet ic acid was cont inuously monitored by digital weighing bal ance which
indicated the extent of react ion. The reaction was stopped with 220 mL of acet ic
acid collected (theoretical 258 mL). React ion yield of the CB oligomer was about
696 g (approximately 94 %). The CB oligomer product, a viscous melt, was ground
into a fine powder. Average molecular weight of oligomer product was determined
following from Guo by rat io of integrated proton NMR spectra between ester and
carboxylic acid end group protons and was within 1% of theoretical molecular
weight. Theoretical molecular weight is determined by subtraction of the number of
moles of acetoxy funct ional groups in the reactor feed mass mult iplied by the mass
of acet ic acid from the molecular weight of the feed monomers t imes their monomer
23
ratio within the feed mass. The CB oligomer product, a viscous melt at 260 °C and
a britt le solid at 23 °C, was ground into a fine powder. An example structure is
shown in Figure 2.2
Fig. 2.2. Example structure of CB oligomer.
For the carboxylic acid end-capped oligomer CB2, 86.8 g TMA, 205.9 g IPA, 372.2
g ABA, and 335.1 g BPDA were used with the same procedure as above for CB
oligomer. 215 mL of acet ic acid was collected after 3 hours at 260 °C (theoretical
272.9 mL) and react ion yield of the AB oligomer was 690.8 g (approximatel y 95
%). Average molecular weight of oligomer product was likewise as for oligomer CB
determined to be within 1% of theoretical molecular weight. The CB2 oligomer
product, a viscous melt at 260°C and a brit tle solid at 23°C, was ground into a fine
powder. An example structure is shown in Figure 2.3.
24
Fig. 2.3. Example structure of CB2 oligomer.
For the acetoxy end-capped oligomer AB, 134.6 g TMA, 346.1 g ABA, and 519.3 g
BPDA were used with the same procedure as above. 195 mL of acet ic acid was
collected after 3 hours at 260 °C (theoret ical 230 mL) and react ion yield of the AB
oligomer was about 731 g (approximately 95 %). Average molecular weight of
oligomer product was determined by ratio of integrated proton NMR spectra
between ester and acetoxy end group protons and was within 1% of theoretical
molecular weight. Theoretical molecular weight is determined by subtract ion of the
number of moles of carboxylic acid functional groups in the reactor feed mass
mult iplied by the mass of acet ic acid from the molecular weight of the feed
monomers t imes their monomer rat io within the feed mass. The AB oligomer
25
product, a viscous melt at 260 °C and a brit tle solid at 23 °C, was ground into a fine
powder. An example structure is shown in Figure 2.4.
Fig. 2.4. Example structure of AB oligomer.
For the acetoxy acid end-capped oligomer AB2, 89.1 g TMA, 70.4 g IPA, 382.0 g
ABA, and 458.5 g BPDA were used with the same procedure as above for CB
oligomer. 203.7 mL of acet ic acid was collected after 3 hours at 260 °C (th eoretical
254.6 mL) and react ion yield of the AB oligomer was 715.7 g (approximately 96
%). Average molecular weight of oligomer product was likewise as for oligomer AB
determined to be within 1% of theoret ical molecular weight. The AB2 oligomer
product, a viscous melt at 260°C and a brit tle solid at 23°C, was ground into a fine
powder. An example structure is shown in Figure 2.5.
26
Fig. 2.5. Example structure of AB2 oligomer.
PTFE was demonstrated to be a potent lubricat ing agent, enabling low wear and
COF when blended with ATSP at low weight percent, therefore we tailored the
ATSP polymer backbone to incorporate fluorinated ethylene units. For this purpose,
perfluorosebacic Acid (PFSA) was used. A1F, C1F, A2F and C2F, which are
analogous to amorphous oligomers A1 and C1 and liquid crystalline oligomers A2
and C2, respect ively, with a small concentration of PFSA (5% weight) were made.
To produce carboxylic acid end-capped fluorinated oligomer C1F, 37.9 g TMA,
12.5 g PFSA, 40.6 g IPA, 97.3 g ABA, and 70.0 g HQDA were mixed in a 500 mL
spherical reactor flask and a react ion procedure analogous to that of CB oligomer
was conducted. The flask was equipped with a three -neck head connected to an inlet
inert gas, a screw-type impeller driven by an overhead mechanical st irrer, a J-type
thermocouple, and an Allihn-type condenser valve to offer a toggle between reflux
and dist illat ion modes. The reactor was cont inuously purged with argon while
emplaced within a Glas-Col aluminum-housed electric heat ing mant le operated via
temperature controller with the thermocouple operat ing as feedback. The reactor
27
was heated to 260 °C for 15 min to obtain a low-viscosity melt during which
st irring was maintained at 300 revolut ions per minute (rpm) and which evidenced a
substant ial acet ic acid by-product which was refluxed during this stage. After
refluxing for 15 minutes, the condenser was toggle to dist illat ion mode and acet ic
acid condensat ion by-product was collected in a 100 mL graduated cylinder and the
mass of the acet ic acid was cont inuously monitored by observat ion of volume,
which indicated the extent of react ion. The react ion was stopped with 64 mL of
acet ic acid collected (theoretical 75.7 mL). React ion yield of the C1F oligomer was
178.9g (approximately 98 %). The C1F oligomer product, a viscous melt at 260 °C
and a britt le solid at 23 °C, was ground into a fine powder. An example of PFSA
incorporation into the backbone is shown in Figure 2.6.
For the carboxylic acid end-capped perfluorinated oligomer C2F, 19.2 g TMA, 56. 5
g IPA, 98.9 g ABA, 71.1 g HQDA, and 12.5 g PFSA were used with the same
procedure as above. 65 mL of acet ic acid was collected at 260 °C (theoretical 76.9
mL) and react ion yield of the C2F oligomer was about 178.7 g (approximately
98%). The C2F oligomer product, a viscous melt at 260 °C and a britt le opaque
solid at 23 °C, was ground into a fine powder.
For the acetoxy end-capped perfluorinated oligomer A1F, 42.5 g TMA, 29.4 g IPA,
36.5 g ABA, 137.5 g HQDA, and 12.5 g PFSA were used with the same proced ure
as above. 61.5 mL of acet ic acid was collected at 260 °C (theoretical 72.8 mL).
28
For the acetoxy end-capped fluorinated oligomer A2F, 21.8 g TMA, 30.2 g IPA,
93.3 g ABA, 100.6 g HQDA, and 12.5 g PFSA were used with the same procedure
as above. 63 mL of acet ic acid was collected at 260 °C (theoretical 74.2 mL).
Reactor vessel configurat ions used for ATSP oligomerizat ions is shown in Figure
2.7.
Fig. 2.6. Backbone incorporation and example react ion between aromatic
const ituent and PFSA.
29
Fig. 2.7 (Left) 500 mL reactor used for aromatic polyester oligomerizat ions. (Right)
2000 mL reactor used for aromatic polyester oligomerizat ions.
2.2: Justification of Oligomer Architecture
Frich, Economy, and subsequent researchers appear to have never explicit ly stated
their rat ionale for some features of the architecture of the oligomers seen in Table
5.1.
30
Table 2.1. Oligomer molar feed rat ios, theoretical number average molecular
weights, funct ionality, and softening points as determined by thermomechanica l
analysis. Highlighted oligomers were used hereafter. From Frich [ 47]
The mole rat io of funct ional groups relat ive to mole rat io of crosslinker necessary
to be in place is for a 3-funct ionality crosslinker (TMA) is defined by funct ional
groups = crosslinker + 2 . Examples are: C1, which would be funct ional groups = 2
+ 2 = 4; C2 funct ional groups = 1 + 2 = 3; etc. For example, if you sum C1’s molar
ratio basis count of COOH (2*3+3*2+6*1=18) and the AcO (4*2 + 6*1 = 14) count
of as COOH (carboxylic acid) is greater than AcO (acetoxy) by 18 - 14 = 4, that
means that the product oligomer will have (on average, since it is a condensat ion
product) four carboxylic acid funct ional caps. A similar process is used for all non -
perfluorinated ATSP oligomers.
31
The funct ional group / crosslinker relat ionship works different ly if you use a
crosslinker with >3 funct ionality. For example, [1,1' -biphenyl]-3,3',5,5'-
tetracarboxylic acid is a 4-funct ional carboxylic acid monomer . For a crosslinker
with a funct ionality of 4, the number of funct ional groups scales as 2*N+2. For an
imaginary crosslinker of funct ionality 5, it would be 3*N+2 and so on. So, the
generalizable rule would be where N is the mole rat io of the crosslinker and C is the
funct ionality count of the crosslinker, the number of funct ional groups (F) needs to
be: F = (C-2)*N+2. A final generalizat ion for arbitrary crosslinker funct ionality and
molar rat io is:
2.3: Theoretical Crosslink Density and Branch Configuration
In addit ional to the funct ional group equimolar approach rat ios used by Frich and
all subsequent workers, it is possible to deliberately bias the resin towards content
of free AcO and COOH funct ional groups. Table 2.2 summarizes number average
theoretical mass between crosslinks (M c)N, branch coefficient xn, funct ional group
ratio COOH/AcO and mass feed rat io for the C1A1 C2A2 oligomer set.
32
Table 2.2. Mass feed rat io and resultant structure parameters.
Sample Oligomer mass feed ratio (C#/A#) Oligomer molar feed ratio (C#/A#) COOH / AcO xn (M̄c)N
ID# (g)
C1A1 1.105 1.000 1.00 0.1440 834.10
C1A2 1.140 1.000 1.33 0.1051 1142.82
C1A2+ 1.243 1.087 1.45 0.1063 1129.92
C1A2- 1.029 0.900 1.20 0.1030 1166.12
C1A2= 0.857 0.750 1.00 0.1010 1189.21
C2A1 1.080 1.000 0.75 0.1100 1117.30
C2A1- 1.224 1.133 0.85 0.1075 1117.30
C2A1= 1.440 1.333 1.00 0.1040 1091.91
C2A1+ 1.584 1.467 1.10 0.1020 1177.55
C2A2 1.117 1.000 1.00 0.0718 1672.84
Where xc equals moles of cross-linked monomeric units divided by moles of monomeric units
present and crosslink density is defined by [48]:
where Mu is the sum of the formula average molecular weights of the oligomers after accounting
for cure loss (number of matched carboxylic acid and acetoxy functional groups x mass of acetic
acid). COOH/AcO is determined by formula basis of functional groups of oligomer feed.
33
CHAPTER 3
DEVELOPMENT OF PRACTICAL COATING SCHEMES UTILIZING ATSP
3.1: Background
The higher performance requirement s and lighter weights in air-condit ioning and
refrigerat ion (ACR) compressors as well as operating condit ion requirements that
have been expanding in terms of higher speed and load has made tribological
characterist ics of interact ing surfaces within the compressor environment play a
significant role in the compressor’s reliability[86]. Compressor reliability, and
generally the interact ion of mechanical systems ut ilizing tribo -pairs, can degrade as
components become more worn [86,87]. This can have secondary issues such as
spikes in power draw. In order to better resist cascading wear caused by hard third -
body wear, an advanced protective coat ing on the interact ing surfaces may be
prudent to mit igate this mechanism in addit ion to the boundary/mixed lubricat ion
that is tradit ionally employed. [117] Due to favorable tribological performance as
well as heavy market ing by private ent it ies owning tradename polymers of this
class, resin-bonded polytetrafluoroethylene (PTFE)- and polyetheretherketone
(PEEK)-based PTFE-filled polymeric coatings have received interest in ACR
compressor applicat ions, as a potential solut ion to supplement and potentially
replace convent ional oil lubricants. However, compared to a great amount of
research and experiments done so far for bulk of polymers , there is limited
literature on the tribological performance of polymeric coat ings.
34
The main advantages of the polymeric-based coatings are their relat ively low cost
and simple substrate surface condit ioning (i.e., no need for expensive surface
preparat ion before coating) as compared to advanced hard coat ings (i.e. DLC, [87]).
However, commercially available polyamideimide, polyimide and
polyetheretherketone-based polymeric coatings st ill exhibit the following problems:
their wear rate is st ill high (compared to hard coatings);
because they rely heavily on the interact ion between the PTFE/PEEK wear
debris/solid lubricant and the substrate for surface protection, in the presence
of lubricant they may become ineffect ive;
the addit ion of hard part icles in these mixtures scratches the c ounterface,
thus creat ing excessive abrasive wear [117];
the wear debris likely contains hard part icles that can damage downstream
machinery via third-body wear;
the highest glass transit ion polymeric electrostatic coat ings available
previous to developments described here are those that based on PEEK [120],
which evidences a glass transit ion of only 143°C [119]
ATSP was invest igated by Zhang and Economy [105] as a potent ial ultra-low wear
polymeric coating system that can mit igate the above problems. We can ident ify
some of the properties of ATSP that make it id eal for this applicat ion, which
includes:
ATSP is designed to be high-temperature stable, with a glass transit ion
temperature ranging up to 310°C (Figure 5.18)
35
Unlike other soft coatings, like PTFE and PEEK, it offers a crosslinked
structure that is more stable under aggressive condit ions.
It has a unique ability among high temperature thermoset polymers to react
with itself (below thermos-oxidat ive degradat ion region) even after curing,
undergoing interchain transesterificat ion react ions (ITR). This has been
proposed to offer unique funct ionality during processing and during use -
allows reincorporation of wear debris into the coat ing surface.
Like the homopolymer p-hydroxybenzoic acid (commercia lly known as
Ekonol®
), ATSP coat ings show potential for plasma spraying due to the
similarity in structure and properties.
ATSP exhibits very good adhesion to different metal substrates. Copper,
titanium, stainless steel and aluminum bonded by ATSP have h igh lap shear
strengths. [11]
The work by Zhang [105] demonstrated excellent performance in some respects, but there were
several features missing in regards to using this roll-on scheme (Figure 3.3) to apply ATSP as an
effective and relevant tribological coating system.
1. The specific coating methodology (rolling on an oligomeric solut ion in NMP)
has the potential to be very non-uniform
2. It is not amenable to tribo-parts of complex geometry.
3. The ATSP coating itself absent a fluoropolymer lubricat ing phase e videnced
a relat ively high coefficient of frict ion. The fluoropolymer lubricat ing phase
was added by Zhang in the form of a sprinkling of Zonyl onto the cured
surface prior to testing. Within state-of-art polymeric coatings, generally the
36
lubricat ing phase is bonded within the resin. However, the author’s personal
knowledge of sample preparat ion difficult ies during that project is that Zhang
and her coworkers did not have a specific method for introducing a
fluoropolymer lubricat ing addit ive phase during preparat ion
4. All samples were cured under vacuum which makes cure cycles for larger
components or large numbers of components a much more difficult
production problem. As one of the arguments for use of polymeric
tribological coat ings is their simplified processing (and thereby lower
expense) relat ive to elast ically hard protective coat ings.
This chapter details the processing experiments and tribological results that demonstrate that
these concerns have been completely overcome. At this point ATSP-based coatings can be scaled
to arbitrary levels of production and applied to parts of complex geometry – including those
where line of sight to deposition device is not feasible.
Three coating techniques are demonstrated and explored: thermal spray coating, solvent-borne
coating, and electrostatic powder coating. Of these electrostatic powder coating is by a wide
margin found to be the preferred method and as a method has been claiming market share within
extant commercial processes [120] from solvent-borne coatings due to their obvious lack of
solvent and therefore ease with which they can meet VOC criteria for both workplace and
finished products. Additionally powder coatings lend themselves easily towards producing a
uniform coating thickness within the sample by virtue that surfaces already covered with powder
screen the electrostatic field that draws powder in the first place: bare metal is more attractive to
37
the oppositely charged powder than that screened by already existing powder layers. The
precipitation polymerization of fully cured ATSP powders was also explored as a pathway
towards thermal spray is also reported here. Finally, tribological results of the various fabrication
schemes are also summarized.
3.2: Synthesis of Crosslinked ATSP Powder with Controlled Particle Size
Cured Powder Synthesis
Ekonol® has been studied extensively with respect to the appropriate particle size and particle
size distribution for plasma spraying. It is well known that the optimum particle size distribution
for plasma-sprayed coatings is in the range of 30-100 μm. It was thought that ATSP, due to its
similarity in structure and properties to Ekonol®, should exhibit a similar behavior.
To synthesize cured ATSP powders, trimesic acid (TMA), hydroquinone diacetate (HQDA),
isophthalic acid (IPA), and 4-acetoxybenzoic acid (ABA) (molar ratio of 4:11:5:8 respectively)
was charged into a 3-neck reactor with Therminol-66, which was continuously purged with
nitrogen. The monomer mixture was stirred using mechanical stirring during the reaction. The
monomers were then refluxed at 270-285 °C for 30 min. The apparatus was switched to acetic
acid removal and the temperature increased to 270 °C. The reaction was carried out at this
temperature until 90% of the theoretical yield of the by-product (acetic acid) was captured. The
temperature was increased to 320-330 °C for the final 5 hrs. The reaction product was then
filtered and washed with acetone and then finally purified using Soxhlet extraction with acetone
for 24 hours. It was also found possible to instead not advance the yield of the by-product to
quite the same degree and allow instead a partially cured structure. The reaction was simply
38
halted once 75% of theoretical yield of acetic acid was reached. All other steps remained
analogous. This partially cured ATSP derived by this route was used for initial thermal spray
trials which appear below.
In a 2L reactor with a total monomer concentration of 0.113g/mL and a final reaction time of 5
hours, it was found that particle size could be well-controlled with stir speed as shown in Table
3.1. Powders were filtered through progressively finer meshes and then weighed and the results
expressed as a weight percentage (wt%) finer than a given mesh size. Increased stir speed
yielded decreased particle size and by this method a yield of over 90% smaller than 105 μm was
achieved. A 300 minute hold time at 320-330°C was as well found to be viable to produce the
desired particle size distribution.
Table 3.1. Particle size distribution as determined by screening.
Figure 3.1 shows a photograph of the ATSP powder material and a scanning electron microscopy
(SEM) image of the material particulates. Objective for powders for thermal spray coatings is
generally assumed to be a globular, rounded morphology.
39
Fig. 3.1. Cured ATSP powder: a) photograph and b) SEM showing individual particulates.
3.3: Preparation of ATSP coatings
Two types of ATSP coat ings, typically blended with polymeric or other addit ives,
were prepared on metal substrates using either compression sintering or solvent
cast ing. Addit ives that were examined include PTFE, polyimide, MoS 2, graphite,
and mullite while the metal substrates were aluminum ( Al390-T6, air condit ioning
compressor), cast iron (G2 Durabar, refrigerat ion compressor) , and 304 stainless
steel. All coatings were deposited
on 1-inch diameter samples for
tribological test ing using a ball-on-
disk configurat ion. A complete
list ing of all ATSP coatings
fabricated at this init ial stage is
shown in Table 3.2.
a)
b)
Fig. 3.2. Cure cycle for ATSP blend
coatings.
40
Method 1: Compression Sintering
The ATSP powder and appropriate addit ive were mixed in a mechanical blender to
ensure uniform dispersion. These powder mixtures were then sintered on the
selected substrates using the cure cycle shown in F igure 3.2. Interchain
transesterificat ion react ion ( ITR) was ut ilized to form a uniform cohesive film and,
when the addit ive is a polymer, an adequate interpenetrating network. A pressure of
400 psi was applied at the start of the cycle and the ent ire cur e react ion carried out
under nitrogen purge.
Good qualit y coatings were produced however reproducibility was low. The coating
thickness, measured using profilometry is 40-60 microns.
Method 2: Solution Casting
A second method, following from Zhang [105] of applying the blends was also
utilized, wherein ATSP oligomers were mixed in 1.1:1 rat io of A1:C1 at 0.5 g/mL
in N-methylpyrrolidinone (NMP) so that the result ing solut ion had a desired
consistency for ease of spreading.
Fig 3.3. Schematic of solut ion cast ing procedure [5].
41
A schematic of the cast ing process is depicted in Figure 3 Many different
permutat ions were tried including pure ATSP, as well as blends of ATSP with
PTFE, graphite, mullite, MoS2, PI and combinat ions thereof (Table 3.2).
Table 3.2. Comprehensive list of all ATSP coatings fabricated including date
produced, addit ives incorporated, and type of substrate.
SUBSTRATE
Al390-T6 G2 Durabar 304 stainless
Pure ATSP Pure ATSP
5% PTFE 5% PTFE
5% Gr + 5% PTFE 15% PTFE
15% Gr + 5%
PTFE
Pure ATSP
5% and 15% PTFE
Compression
Sintered
5% and 15% Gr
Solution Cast 5% PTFE + 15% Gr
5% PTFE + 5% Gr
15% PTFE + 5% Gr
15% PTFE +5% Gr
5% PTFE
Pure ATSP Pure ATSP
5% and 15% mullite 5% and 15% mullite
5% and 15% PTFE 5% PTFE + 5% mullite
5% mullite + 5% PTFE 5% PTFE + 15%
mullite 15% mullite +
5%PTFE
1% PTFE
1% PTFE 5% PI
5% Gr + 5% PTFE 5% PI + 5% PTFE
15% PI + 5% PTFE
5%PI+5%mullite+5%P
TFE 5% MoS2 1% and 5% MoS2
5% MoS2 + 5% PTFE 1% MoS2 + 1% PTFE
1% PI + 1% PTFE 5% MoS2 + 5% PTFE
5% PI + 5% PTFE 5% PI
1% PI + 5% PTFE
5% PI + 5% PTFE
42
3.4: Tribological Experiments
Since a large number of coat ings were made using three different methods, namely
(a) compression sintering (CS), (b) solut ion cast ing (SC), and (c) thermal spray
(TS), we decided to perform controlled ball-on-disk experiments to screen these
coatings. Such tests would readily provide coat ing adherence to the substrate,
frict ion, and wear performance. Three different condit io ns were used, ident ified as
“mild”, “intermediate” and “aggressive” condit ions, with the majority of tests
performed at the intermediate condit ions. Realist ic compressor condit ions, refer to
higher speeds, but typically lower contact pressures as nominal flat (pin) on flat
contact are encountered, e.g., in a scroll compressor.
Below is a photograph of the tribometer utilizing a ball-on-disk geometry, located
in the tribology laboratory at the UIUC, that was used for all experiments in this
subsect ion (Fig. 3.4). Also, shown is a photograph of a typical disk sample that
carries the coat ing and a chrome steel sphere as the counter surface. A first set of
experiments was performed at a sliding speed of 120 mm/s, a normal load of 5 N
(corresponding to a contact Hertzian pressure of 75.2 MPa) for 30 min (216 m
sliding distance) at ambient laboratory condit ions (designated as mild test ing
condit ions). An intermediate case increased the load to 10 N but decreased the
sliding speed to 60 mm/s (108 m sliding distance) while a final severe case ut ilized
a normal load of 10N (corresponding to a contact Hertzian pressure of 124.2 MPa)
43
for 30 min and returned the sliding speed to 120 mm/s.
Fig. 3.4. Photographs of tribology equipment at UIUC along with standard 1 -inch
diameter sample disk and ¼ inch diameter chrome steel sphere.
The best performing frict ion and wear data were obtained with the 5% PTFE
samples using both cast iron and Al390-T6 substrates. Figure 2 shows a typical
result from the Al390-T6 aluminum substrate, exhibit ing “zero” wear and very
stable coefficient of frict ion. Similar results were obtained on the G2 cast iron
sample as well. The experiments were also repeatable, confirming the excellent
performance of these coatings. These results are extremely promising as they show
superiority over state-of-the-art commercially available polymeric-based coat ings.
Case in point a commercially available state-of-the art PTFE/Pyrolidone coating
was tested under ident ical condit ions and the results were ver y similar, namely
44
stable/similar frict ion coefficient around 0.2 and mild burnishing -no significant
wear. The statement about superiority is that it is expected as shown in earlier data
[5] that at higher stress levels, the ATSP coating showed significant ly lower wear
performance. At the more aggressive wear condit ions, our CS coat ings tended to
flake off, indicat ing poor adherence to the substrates and not poor tribological
performance. Higher percent of PTFE in the coat ing (15%) showed poorer
performance. As seen before, which is somewhat counterintuit ive, pure ATSP
coatings exhibited high frict ion coefficient (0.5 and higher), wearing the chrome
steel ball and yet showed no wear of the coating.
Fig. 3.5. 5% PTFE-blended ATSP compression sintered (CS) onto Al under mild
testing condit ions. a) Graph showing stable COF over t ime (inset is picture of
sample disk), b) Profilometry of sample after wear test showing no material has
been removed compared to baseline.
Solut ions of ATSP were cast and overall, they resulted in more uniform coat ings
and stronger adherence to the substrates since no flaking of the coat ings was
a) b)
45
observed. The frict ion coefficient was similar to the compression sintered coat ings,
with somewhat higher wear rate in the case of 5% PTFE. Under the intermediate
testing condit ions as described above, the cast iron substrates were init ially
examined and the coatings of pure ATSP, 1% PTFE, 5% Mullite, 1%PI+1%PTFE,
and 5%PI+5%PTFE failed prematurely. However, the best performing coat i ngs
were 5% MoS2 and the 5% PTFE+5% MoS2 with wear rates of ca. 9.2 x 10-4
and 4.6
x 10-4
mm3/Nm, respect ively. The frict ion coefficient was steady and the lowest
achieved with the 5% MoS2 coating and steady COF of 0.15 while the second best
is very close with 5%PTFE+5% MoS2. Figures 3.8 and 3.9 display the overall
results of all wear tests and the corresponding coefficients of frict ion. In general,
the wear rates are very low except when mullite is the only addit ive, while the
presence of any mullite results in a sharp increase of the COF.
0 10 20 300.0
0.2
0.4
0.6
CO
F
Time (min) 0 2 4
-40
-20
0
20
40
Heig
ht
(µm
)
Distance (mm)
Test 23_1
Fig. 3.6. 5% MoS2-blended ATSP (CS) and tested at 124.2 MPa Hertzian pressure.
In addit ion, several disks that were coated on 304 stainless steel were tested at the
intermediate wear condit ions. In general, the wear data was not as high quality as
46
the cast iron samples and the COF was higher. The reason for the poorer
performance of the stainless-steel substrates is that it is well known that cast iron
possess good tribological properties and since these coatings are very thin any
penetration of the coat ings will result in exposure of the substrate. Further
invest igat ion as to which substrates are suitable with ATSP, and whether any
optimizat ion of the coatings could be done to improve performance with sub strates
such as stainless steel is needed.
0 2 4-40
-20
0
20
40
Distance (mm)
Heig
ht (
m)
Test14_1
0 10 20 30
0.0
0.4
0.8
1.2
CO
F
Time (min)
Fig. 3.7. 5% PTFE+5% Mullite blended with ATSP (CS) and tested at 124.2 MPa
Hertzian pressure.
A comment is in order in regards to the performance of mullite in the ATSP
coatings, which was init ially unexpected. The fr ict ion coefficient is clearly high (>
0.5), it wears significant ly the chrome steel ball and yet there is no measurable wear
47
of the coat ing. Instead there is material pile up as shown in Figure 3.7. Such
material could be further explored in applications where a higher COF and no wear
are required such as clutches and brakes.
3.5: Preliminary Thermal Spray Coatings
The primary object ive for this task was to thermal spray ATSP powder onto button
coupons for characterizat ion. The goal was to generate fully coalesced coatings
that exhibit minimal oxidat ion/degradat ion. In addit ion, key processing parameters
were ident ified for future work that will yield improvements in coating deposit ion
and performance.
This work was carried out with Prof. Sanjay Sampath at the State University of New
York at Stony Brook in the Center for Thermal Spray Research (CTSR).
Technical Approach
Two thermal spray techniques were selected and used for deposit ion trials; namely,
combust ion spray processing and DC plasma spray processing. The combust ion
spray process entailed using a Terodyn 3000 gun (Messer Eutectic) to deposit the
partially cured ATSP powder on 09/23/11 and the fully cured ATSP powder on
12/21/11-12/23/11. A Sulzer Metco 3MB Plasma torch with a standard 7MC
console was used to generate plasma sprayed coatings on 12/22/11.
48
Fig. 3.8. Wear rates at intermediate test ing condit ions for solut ion cast ATSP
samples. Inset shows a representat ive disk with a dist inct ive wear track and the
subsequent profilometry graph used to calculate wear rate.
49
Fig. 3.9. Average COF at intermediate testing condit ions for solut ion cast ATSP
samples.
A low temperature nozzle (LT 250) was used with the Terodyn combust ion gun
based upon prior experience with spraying polymers. A low temperature shroud
assembly was also used to cool the flame and provide a spray parameter to control
the extent of particle melt ing. It has been determined that the shroud is integral to
prevent polymer degradat ion in flight and can be used to vary the flame
temperature. The Plasma Technique Twin 10 feeder was used to feed the ATSP
powder into the oxy/acetylene flame.
A schematic of the combust ion spray process is shown below in Fig. 3.8. Oxygen
and acetylene are used for combust ion. Nit rogen is used as the carrier gas to inject
50
the powder into the flame. Compressed air is used in the shroud assembly. The
spray gun was mounted on a computer controlled, linear x-y traverse system. A
robot was not required since only flat coupons were sprayed requiring 2-axis of
motion. The traverse speed and standoff distance (distance from nozzle to sample)
can be varied systematically.
Nitrogen
Powder
Feeder
To Air
Compressor
Air Control
Acetylene
Oxygen
Air Shroud
Assembly
NozzleBurning
Gases
Spray
Stream
Acetylene/Oxygen
Flow Control
Nitrogen
Powder
Feeder
To Air
Compressor
Air Control
Acetylene
Oxygen
Air Shroud
Assembly
NozzleBurning
Gases
Spray
Stream
Acetylene/Oxygen
Flow Control
Fig. 3.10. Schematic of Eutectic Terodyn 3000 combust ion spray gun.
The ATSP team evaluated the full range of parameters that are typically optimized
while carrying out thermal spray studies and down-selected the first order
parameters that should be studied in the first series of trials (Figure 11 outlines the
first order parameters that were studied).
At the beginning of the trial day, init ial studies were carried out to demonstrate that
the ATSP powder could melt on a substrate surface. Init ial process parameters were
determined to be 275 and 325 °C by virtue of available DMA profiles at the t ime
51
[20] which suggested a potent ially low enough storage modulus at that range [94] .
Finally all metal substrates were suitably primed using surface grit blast ing with
#24 mesh alumina.
Fig. 3.11. First order parameters (outlined in red) selected to be examined in this
study. Original figure from [121] .
First Trial: Partially-Cured ATSP Powder
Partially-cured ATSP powder derived by solut ion polymerizat ion in Therminol 66
was loaded into the Twin 10 feeder hopper and used for combust ion spray trials.
An init ial set of parameters was chosen based on prior experience with deposit ing
liquid crystal polymer powders. The first two samples in Table 3.3 were for the
purpose of generat ing splats. However, very litt le in -flight melt ing occurred and
the substrate pre-heat temperature was insufficient to promote splatting. Attent ion
was then focused on generat ing coat ings whereby addit ional passes of the spray
torch are needed to heat the polymer during deposit ion. The following parameters
were chosen and then varied as shown in Table 3.3 to influence flame temperature,
52
particle velocity, and substrate temperature. High carrier flow rates were also used
to create high shear upon impact. Temperatures of the deposited coat ing were
recorded using an infrared pyrometer with e missivity set to 0.95.
Spray Distance: 3 inches
Air Back Pressure: 60 psi
Acetylene pressure/flow: 15 psi / 15 FMR
Oxygen pressure/flow 50 psi/13 FMR
Argon Carrier pressure/flow: 4 bar / 10 FMR
Traverse Rate: 100% (2 inches/sec)
Table 3.3. Partially-cured ATSP sample summary (combust ion spraying).
Sample
Number
Pre-
Heat
(°F)
Air
Shroud
(psi)
Cycle
s
Powder
Wheel
RPM %
Coating
Temp
(°F)
Thickne
ss (mils)
Comment
0923-01 110 60 1 5 NA. Splats No splats
0923-02 270 60 1 5 NA. Splats No splats
0923-03 250 60 3 30 450 10 Poor melting
0923-04 325 60 2 30 550 7 Some melting
0923-05 320 50 2 30 420 6.5 Poor melting
0923-06 320 45 2 30 500 6 Poor melting
0923-07 450 45 1 30 520 3 150%
Traverse Speed. Some
Melting
0923-08 450 45 3 30 520 7 150%
traverse.
0923-09 550 45 2 30 600 5.5 Some melting
0923-10 550 20 1 30 >600? 2 Degraded
0923-11 650 None 1 30 380 1.5 No Flame –
Powder Only. Coating is
WHITE.
Post-heated to 450F, coating
turned
BLACK
0923-12 600 50 2 30 580 5 Some Melting
53
The part ially cured ATSP did not exhibit melt flow behavior like a tradit ional
thermoplast ic polymer. Coatings were not completely coalesced although some
particle melt ing was observed under a stereomicroscope. At low temperatures
(clear coatings), the ATSP powder does not melt uniformly, leading to a coating
that is not fully reacted. The polymer is sensit ive to thermal -oxidat ion as witnessed
by the discoloration at coating temperatures above 500F. It was difficult preheat ing
the substrate to temps above 450F and keeping the substrate at that temperature just
prior to deposit ion due to substrate cooling effects. These first order parameters
produced a wide variety of results, indicating that they are indeed the crit ical
parameters to optimize, and also providing a processing window to be further
refined. Figure 3.12 shows pictures of the 12 substrates after thermal spraying.
Fig. 3.12. Twelve coated substrates produced by varying the first order parameters.
Several of these samples were also tested under the int ermediate wear condit ions.
Overall, they had a very high level of COF (as expected for pure ATSP coatings)
and the two samples on stainless steel (0923-3 and -6) wore away quickly and the
test had to be stopped prematurely. Sample 0923-11 on cast iron appeared to
54
perform better Test 30 is shown below: The tribological performance is as expected
or better than earlier pure ATSP specimens. An addit ional qualitat ive assessment is
that the coatings did not peel off by attempts to delaminate via human finger n ail,
indicat ing that adhesive strength was
0 10 20 300.0
0.4
0.8
1.2
CO
F
Time (min) 0 2 4
-40
-20
0
20
40
He
igh
t (µ
m)
Distance (mm)
Test 30_2
Fig. 3.13. Pure ATSP (TS) on cast iron tested at 124.2 MPa Hertzian pressure.
Second Trial: Fully-Cured ATSP Powder
The fully-cured powder was chosen for the 2nd
round of spray trials. The powder
was also passed through a ~60 mesh sieve to screen out any large polymer part icles.
Based on the init ial spray trials conducted with the part ially cured ATSP powder, a
resist ive substrate heater was implemented to maintain the high substrate
temperature throughout the deposit ion process. The heater employs a controller
based on a Type K TC input as feedback to controls the heater power. It was
determined that the maximum heater setpoint was 500 °C (932F). A range of
substrate heat ing temperatures was chosen to determine optimal heat ing to melt the
polymer powder. The following parameters were fixed:
55
Spray Distance: 3 inches
Air Back Pressure: 45 psi
Acetylene pressure/flow: 15 psi / 15 FMR
Oxygen pressure/flow 50 psi/13 FMR
Argon Carrier pressure/flow: 4 bar / 10 FMR
Traverse Rate: 100% (2 inches/sec)
Powder Wheel RPM rate: 30%
Table 3.4. Fully-cured ATSP sample summary (combust ion spraying).
Sample
Number
Heater
Setpoint
(°F)
Button
Surface
Temperatur
e (°F)
Cycle
s
Thicknes
s (mils)
Comment
1222-01 none 70 2 17 Unmelted, color = brown
1222-02 392 212 2 14 Unmelted, color = brown
1222-03 662 302 2 9 Unmelted, color = brown
1222-04 932 482 2 5 Some melting, color =
brown
1222-05 932 572 2 3 Some melting, color =
black
1222-06 932 572 1* 3 Some melting, color =
black
*50% traverse speed
The tribology of these are shown in Figure 3.14 tests 35 (sample 03) test 36 (sample
04), test 37 (sample 07), test 38 (sample (08) and clearly Sample 4 performing the
best, as it did not wear the coating, despite the high frict ion and the ball wear
56
The fully-cured ATSP powder also did not exhibit tradit ional melt flow behavior
and became dark in color due to thermal oxidat ion before melt ing occurred
(comparing samples -04 to -05). The coat ing surface is rougher than coatings made
by the part ially-cured ATSP powder due to the larger powder size.
Next, a Sulzer Metco 3MB Plasma gun was used to examine the feasibility o f
deposit ing fully-cured ATSP polymer powder. A number of settings were used and
2 buttons were generated from the following parameters.
1222-07: GH nozzle, #2 powder port, 500 Amps/70 V, 80F Ar Flow, 15F
Hydrogen Flow.
6 passes = 1 mil thickness. Coat ing is black. Very litt le deposit ion,
most of the powder was not melted and degraded in-flight.
1222-08: GH nozzle, #1 powder port down stream injector, 400 Amps/70 V, 80F
Ar Flow, 5F Hydrogen Flow.
10 passes = 2 mil thickness. Coat ing is black. Very litt le deposit ion,
most of the powder was not melted and degraded in-flight.
The plasma spray trials were not successful in melt ing the powder and caused
extensive oxidat ion/degradat ion to the ATSP powder based on the parameters
chosen during that experiment . This is thought to be a consequence of the highly
crosslinked cured C1A1 structure that prevented melt ing during thermal spray and
therefore appropriate splat format ion [93].
57
Fig. 3.14. Wear performance of several TS ATSP specimens tested at 124.2 MPa
Hertzian pressure.
3.6: ATSP Thermal Spray Coatings Derived By Two-Part Oligomeric Approach
To address the deposit ion limitat ions encountered in 3.5 and improve the thermal
window between melt ing point and degradat ion temperature, an alternat ive
synthet ic approach was employed. As ATSP material derived by oligomeric route
was discovered during this course of development to have a superior
thermogravimetric curve (Figure 3.15) than those intended to be cured powders.
Powders of C2A2 and CBAB were part ially cured at 230 °C and 260 °C to produce
int imate blending and advance molecular weight while retaining greater melt flow
during processing. The powders were designated as follows: CBAB -230, CBAB-
58
260, C2A2-230 and C2A2-260. Target materials were ground and wet sieved into
the desired particle size range. Figure 3.15 shows the TGA analysis of the produced
ATSP powders.
Fig. 3.15. (Top Left) Example thermogravimetric trace with temperature versus
time in red and oligomer-derived C2A2 in blue. (Top Right) Comparison of melt
oligomerizat ion derived C2A2 with several condit ions of cured powder synthesis
under both flowing air and N2 exposed to temperature profile at right . Powders
produced via solut ion-suspension route are designated “SR”.
It can readily be observed from Figure 3.15 that powders produced via the route
used by Huang [107] and Zhang [19,20] evidence an addit ional 10-15% weight loss
during a 371°C hold experiment.
A Beckman Coulter Particle Size Analyzer was used to measure the powder size
distribut ion for each powder, as summarized in 3.5. The powders were all of
similar size distribut ion and quite flowable and, thus, did not require a flow
enhancer to be blended such as cabosil (silica fume).
59
Table 3.5. Powder size distribution summary.
ATSP Powder
Designation
Mean
(microns)
d10
(microns)
d90
(microns)
CBAB-230 72 36 110
CBAB-260 71 36 109
C2A2-230 65 29 105
C2A2-260 67 34 102
Mild steel coupons measuring 1”x2”x0.125” were used as substrates in the current
study. All coupons were grit -blasted using 60 mesh aluminum oxide grit at 60 psi.
Fixturing was set up to enable one coupon to be coated at a t ime using a magnet ic
substrate holder.
A Messer-Eutectic Terodyn 3000 gun with LT250 nozzle and low temperature
shroud assembly was used to deposit the four polymer powders. It has been
determined that the shroud is integral to prevent po lymer degradat ion in flight and
can be used to vary the flame temperature. Oxygen and acetylene are used for
combust ion. Argon was used as the carrier gas to in ject the powder into the flame.
The spray gun was mounted on a computer controlled, liner x-y traverse system.
An init ial set of parameters was chosen based on prior experience with deposit ing
the copolyester powders from 3.4. Both the gun traverse speed and number of spray
cycles (sets of passes) was varied to influence coating temperature. Init i ally the
intent was to vary substrate preheat temperature as well but it became evident that
the supplied polymers did not require a preheat to induce melt flow. However, a
butane torch was used to slight ly preheat the samples to 70°C where moisture was
60
visibly removed from the steel substrate surface and coat ings were then deposited
once the steel surface measured 35°C. Surface temperatures of the steel and
deposited polymer coatings were measured with a simple hand -held infrared
pyrometer with emissivit y set to 0.95. Pictures of the thermal spray coating
deposit ion are shown in Fig. 3.16.
Spray Distance: 3 inches
Air Back Pressure: 60 psi
Acetylene pressure/flow: 15 psi / 15 FMR
Oxygen pressure/flow 50 psi/13 FMR
Argon Carrier pressure/flow: 4 bar / 10 FMR
Air Shroud Pressure: 45 psi
Powder Wheel RPM: 15%
Gun Traverse Rate: F Value = 500 Table 3.6. Thermal spray coating summary with process parameters and resultant coating status.
Sample
Number
Pre-
Hea
t
(°C)
Traverse
Rate, (%)
No. of
Cycle
s
Coating
Temperature
(°C)
Thickness
(mils)
Comment
CBAB-230-0 Non
e
50 1 150 4 Practice sample – not grit-blasted.
Insufficient melting
CBAB-230-1 100 50 1 210 4 Too Hot, visibly degraded
CBAB-230-2 35 50 1 175 3 Mostly coalesced
CBAB-230-3 35 50 2 200 10 Too hot, visibly degraded
CBAB-230-4 35 70 2 170 5 Mostly coalesced, rough
CBAB-230-5 35 70 2 180 5 Repeat of #4
CBAB-230-6 35 90 3 180 7 Rougher, not fully coalesced
CBAB-260-7 35 70 2 160 9 Rougher, not fully coalesced
CBAB-260-8 35 60 2 170 11 Rougher, not fully coalesced
CBAB-260-9 35 50 1 175 7 Mostly coalesced
CBAB-260-10 35 50 1 170 6 Repeat of #9
C2A2-230-11 35 70 2 200 7 Fully Coalesced, glossy
C2A2-230-12 35 70 2 190 7 Repeat of #11
C2A2-230-13 35 50 1 165 4.5 Fully Coalesced, glossy
C2A2-230-14 35 50 1 170 4.5 Repeat of #13
C2A2-230-15 35 50 1 170 4.5 Repeat of #13
C2A2-260-16 35 50 1 165 5 Fully Coalesced, glossy
C2A2-260-17 35 50 1 170 5 Repeat of #16
C2A2-260-18 35 50 1 170 4.5 Repeat of #16
C2A2-260-19 35 50 1 160 5 Repeat of #16
C2A2-260-20 35 70 2 175 7 Fully Coalesced, glossy
61
Fig. 3.16. Thermal spray deposition of ATSP copolyester powder onto steel coupons.
The CBAB-230 was the first powder used for establishing gun traverse rate and
substrate preheat parameters. It was quickly observed that the new ATSP supplied
powder had great ly improved melt flow compared to the powders used during the
init ial trials in 3.4. The powders did not require a pre-heat to induce melt flow.
When the substrates were preheated to 100C, the coating got too hot as evident from
the discoloration and visible degradat ion. A small processing window exists
whereby coat ings deposited around 175C were not fully coalesced yet coatings
deposited around 200C were thermally degraded without being completely
coalesced. It should be noted that the coating temperatures recorded by the infrared
62
pyrometer are approximate since the coating cools rapidly upon deposit ion but an
attempt was made to record the coat ing temperature in a consistent manner
immediately upon complet ion of the last spray pas. Traverse speed and number of
spray cycles were varied to improve melt ing and maintain a thickness ~ 125 μm.
The CBAB-260 exhibited slight ly less melt flow than the CBAB-230 powder. The
same parameters for sample number CBAB260-7 were used as CBAB230-4
however, the CBAB260-7 was slight ly rougher and less melted.
The C2A2-230 powder was deposited using the same 70% speed/2 cycles as noted
above and resulted in a completely coalesced, glossy coat ing. Further parameter
optimizat ion resulted in decreasing the speed to 50% and deposit ing the coat ing in 1
cycle of passes, yielding a thickness of approximately 125 μms as measured with a
micrometer. Coating temperatures measured ~ 170-180°C for the fully coalesced,
glossy coatings. Similar results were also obtained for the C2A2 -260 powder,
yielding a fully coalesced coat ing.
3.7: Refinement of solvent-borne coating deposition
As one possible pathway towards practical deposit ion of ATSP coatings on
inorganic substrates for tribologically-relevant polymeric coat ings, solvent -borne
deposit ion was explored more systematically after init ial posit ive results shown
above in 3.2 and 3.3 although it was necessary to greatly refine the process. It was
63
known in several instances of prior literature [43,51,105] that ATSP oligomers may
be dissolved in NMP (n-methylpyrrolidone) and this was employed above. However,
it was not clear that this was the ideal solvent for every situat ion. A range of
available laboratory solvents were evaluated to determine which could solvate
ATSP oligomers via a screening trial at 0.15 g/mL. For solvent -borne coat ings, a
higher concentrat ion is better for the necessity of building a sufficient ly thick dry
coating layer. Deposit ion at lower concentration would require a very thick layer of
solut ion which would preclude effect ive deposit ion on complex parts or vertical
surface.
Solvents were acquired from Alfa Aesar and used without further treatment. C1A1
and C2A2 oligomer sets were prepared at a 1.1:1 mass-basis carboxylic acid-capped
to acetoxy capped rat io. 1.5 grams of mixed oligomers were added to 10mL of
solvents in a PTFE-capped 20mL scint illat ion vial and st irred by magnet ic st irring
either at 70°C or at 10°C below their boiling point (whichever was lowest) for 4
hours. The results are summarized in Table 3.7. Generally it was found that aprotic
polar solvents were compat ible with ATSP oligomers with 1,4-dioxane providing
the only posit ive except ion.
One obvious defect that was observed in early coat ings of ATSP derived from the
examined solvents was that low boiling point solvents tended to readily produce
blisters on films of ATSP especially at higher oligomer loading concentrat ions.
Solut ions of 0.15g/mL C1A1 in compat ible solvents were deposited onto articles o f
64
scrap aluminum 6061 and cured under vacuum with a temperature cycle ramping
1°C/min to 330°C. Coatings were subsequent ly examined for blistering, which is
the format ion of either raised bubbles or depressions on a polymer coat ing.
Blistering was found to decrease with increasing solvent boiling point with solvents
that boil below 160°C generally found to cause blistering in ATSP coatings
regardless of flash off process in coat ings thicker than 5 μm. This is illustrated in
Figure 3.17, with decreasing blistering associated with increasing solvent boiling
point. Addit ionally, solut ions of arbitrary concentration were found to be capable o f
producing blister-free coat ings given a long enough flash off, however a blister -free
flash off in less than 2 hours required either a solut ion concentration of <0.30 g/mL
or a thickness lower than is typically found in tribological polymer coatings.
Fig 3.17. Coatings of 0.15 g/mL C1A1 oligomers in solvent. Solvent used from left
to right: tetrahydrofuran, 1,4-dioxane, dimethylformamide, n-methylpyrrolidone in
ascending boiling from left to right .
65
As described above, a wide assortment of reinforcing and lubricat ing are generally
used in modern solid lubricant coatings. To 10 mL of dimethylformamide, dimethyl
was added 0.075 g of available addit ives (equivalent to 5wt% of 0.15g/mL ATSP
solut ions commonly used in solve spray process). These included bo th DuPont’s
Zonyl® line of PTFE and select inorganic addit ives (aluminum silicate,
molybdenum disulfide, and graphite) . Results are summarized in Table 3.8. The 20
mL scint illat ion vials containing the solvent and addit ive were then shaken
vigorously by hand for 60s and interact ion of solvent and addit ive observed by eye.
The Zonyl line of addit ives was chosen due to its small particle size in comparison
to the Teflon® part icle size ranges. A table of relevant properties of Zonyl based
coatings is available in [109].
It was found that inorganic addit ives sett led at the bottom of the scint illat ion vial in
general conformation to Stoke’s law on a time -basis by means of a stop watch as
measured by eye and evidenced wetting in all of the three down-selected solvents.
With guidance from tribological results above indicat ing that 5wt% concentration of
PTFE resulted in the lowest COF and wear rate, solut ions ut ilizing Zonyl MP1100
as it produced the lowest degree of agglomerat ion and dispersed well in high
boiling point polar aprotic solvents. A proposed mechanism for this dist inct ion is
that Zonyl MP1100 and several others demonstrating at least mediocre dispersion
within NMP are actually produced by first polymerizing a full molecular weight (1 -
4 * 106 g/mol) grade of PTFE which is then irradiated by electron beam under either
an air or CO2 atmosphere [109] which results in the presence of carboxylic acids as
66
funct ional end caps. This also significant ly reduces the molecular weight of the
product down to 0.5 * 106 g/mol. The bare presence of the COOH is thought to give
it significant ly more interact ion with polar solvents in comparison to full MW PTFE
examined such as Saint -Gobain’s Norton® grade with a MW of 1*106 g/mol MW.
Solvated oligomer viscosity of neat oligomers or 1.1:1 matched sets in NMP was
conducted a Cole-Parmer 98936 spindle-type viscometer with a LV-4 spindle at 60
rpm. Results are summarized in Figure 3.19. It was found that the COOH -capped
oligomers created an increase in viscosity relat ive to the acetoxy-capped oligomers
of similar configurat ion. This is likely straightforwardly due to the hydrogen
bonding afforded by the carboxylic acids. An addit ional curious effect was found in
the far higher viscosity of the C2 oligomer at 0.30 g/mL as compare d to the C1
oligomer at the same concentrat ion and a non-linear temperature dependence. The
C2 oligomer actually possesses far less hydrogen bonding per unit mass (number of
carboxylic acids per number average molecular weight – 0.0016 g-1
) versus the
0.0021 g-1
for C1. Addit ionally, the softening point determined by Frich [ 48] for C2
is nearly 30°C lower than that of C1. A clue might be found in the similar
temperature dependence, though quant itatively lower viscosity found in A2. Both
C2 and A2 had previously been found to evidence a nematic liquid crystalline
phase. Unfortunately, a detailed causal mechanism has not been determined for this
pattern of behavior. Another interest ing phenomena is the viscosity profile of the
mixed oligomers at 0.30 g/mol wit h temperature. As can be seen in the lower plots
of Figure 3.19, the viscosity of C2A2 now largely loses its temperature dependence
67
below 50°C while the same condit ions for C1A1 now has an ent irely linear (with
temperature) viscosity profile. A more detailed invest igat ion with more variances in
temperature, concentration, oligomer MW, COOH-concentration, neat un-solvated
oligomer rheology, and LC-phase behavior would be necessary to understand the
underlying mechanism.
At this point, trial coatings of ATSP were employed using an industrially relevant
deposit ion device, namely a high volume/low pressure (HVLP) paint spray gun. The
specific device was a Central Pneumatic 46718 HVLP touch up spray gun with a
120 mL top-fed cup size and a 0.8 mm adjustable nozz le.
Specimen substrates were first grit -blasted via 60 psi of compressed air carrying 40
mesh (400 μm) alumina grit. Specimen substrate for tribological testing were
generally gray cast iron G2 Durabar in a 3” diameter disk with 4 beveled screw
holes for implementat ion into a tribotest platform which will be discussed in sect ion
3.9.
68
Table 3.7. Summary of ATSP oligomer solubility in assorted laboratory solvents. Boiling point
(BP), density, dielectric constant and dipole moment obtained from [92].
Solvent Structure Class BP (K) Density (g/mL) Dielectric constant Dipole moment (D) Miscible C1A1 to 0.15 g/mL Miscible C2A2 to 0.15 g/mL
acetic acid Polar, Protic 391 1.049 6.20 1.74
acetone Polar, Aprotic 329 0.791 21.00 2.88
chloroform Non-polar 334 1.483 4.81 1.04
1,2-dichloroethane Non-polar 357 1.253 10.50 1.80
dimethylformamide Polar, Aprotic 425 0.948 38.00 3.86
dimethylacetamide Polar, Aprotic 438 0.937 37.78 3.72
dimethylsulfoxide Polar, Aprotic 462 1.100 48.00 3.96
1,4-dioxane Non-polar 374 1.033 2.30 0.45
ethanol Polar, Protic 351 0.789 24.55 1.69
ethyl acetate Polar, Protic 350 0.894 6.02 1.78
methyl ethyl ketone Non-polar 353 0.805 18.51 2.76
n-methylpyrrolidone Polar, Aprotic 475 1.028 32.20 4.09
pyridine Polar, Aprotic 388 0.982 12.40 2.69
tetrahydrofuran Polar, Aprotic 339 0.886 7.50 1.75
xylenes Non-polar ~411 0.860 2.2-2.6 0.07-0.45
toluene Non-polar 384 0.867 2.38 0.36
water Polar, Protic 373 1.000 80.00 1.85
69
Table 3.8. Summary of additive suspensions in DMF, DMAc, and NMP. Suspensions were
characterized based on tendency to accumulate at meniscus (↑) or bottom of scintillation vial (↓)
and the degree of agglomeration was rated 0 to 5 with 0 being no visible macro agglom erates, 3
being a tendency towards a clear solution and up to mm -scale agglomerates, and 5 being instant
clearing and up to cm-scale agglomerates.
Additive Accumulate ↑↓ Agglomerates (0-5) Accumulate ↑↓ Agglomerates (0-5) Accumulate ↑↓ Agglomerates (0-5)
Zonyl TE5069AN ↑ 2 ↑↓ 2 ↑ 3
Zonyl MP1000 ↑ 4 ↓ 2 ↑ 4
Zonyl MP1100 ↓ 1 ↓ 0 ↓ 2
Zonyl MP1200SZ ↑ 4 ↑ 4 ↑ 4
Zonyl MP1300 ↑ 4 ↑ 4 ↑ 4
Zonyl MP1500 ↑ 5 ↑ 5 ↑ 5
aluminum silicate ↓ 0 ↓ 0 ↓ 0
MoS2 (<2μm) ↓ 0 ↓ 0 ↓ 0
graphite (44μm) ↓ 0 ↓ 0 ↓ 0
DMF DMAc NMP
70
20
30
40
50
60
70
800.000.05
0.100 .1 5
0 .2 00 .2 5
0 .3 00
20
40
60
80
100
120
140
160
180D
ynam
ic V
iscosity
(cP
)
Temperature (°C)
Concentration (g/mL)
20
30
40
50
60
70
800.00
0.05
0 .10
0 .15
0 .2 00 .2 5
0 .3 00
200
400
600
800
1000
1200
1400
Dyn
am
ic V
iscosity
(cP
)
Temperature (°C)
Concentration (g/mL)
20
30
40
50
60
70
800.00
0.05
0 .10
0 .1 5
0 .2 00 .2 5
0 .3 00
20
40
60
80
100
120
140
160
180
Dyn
am
ic V
iscosity
(cP
)
Temperature (°C)
Concentration (g/mL)
20
30
40
50
60
70
800.00
0.05
0 .10
0 .15
0 .2 0
0 .2 5
0 .3 00
20
40
60
80
100
120
140
160
180
Dyn
am
ic V
iscosity
(cP
)
Temperature (°C)
Concentration (g/mL)
20
30
40
50
60
70
800.00
0.050 .10
0 .1 50 .2 0
0 .2 5
0 .3 00
20
40
60
80
100
120
140
160
180
Dyn
am
ic V
iscosity
(cP
)
Temperature (°C)
Concentration (g/mL)
2030
4050
6070
800.00
0.050 .10
0 .1 50 .2 0
0 .2 50 .3 00
20
40
60
80
100
120
140
160
180
Dyn
am
ic V
iscosity
(cP
)
Temperature (°C)
Concentration (g/mL)
Fig. 3.18. Oligomer viscosit ies by spindle -type viscometer with temperature and
concentration. (Top Left) C1, (Top Right) C2, (Middle Left) C1, (Middle Right) A2,
(Bottom Left) C1A1, (Bottom Right) C2A2.
71
Figure 3.19 demonstrates what happens if a PTFE grade that does not disperse well
in its solvent versus adequate dispersion. 0.30 g/mL of various ATSP
oligomer/5wt% MP1100 in NMP generally remained dispersed indefinitely with
mixture solidificat ion occurring on the time-scale of weeks being the main
mechanism of non-performance. All solvent -borne coating was generally conducted
at 0.30 g/mL with the mixture warmed to 70°C. Figure 3.20 demonstrates an
example roughness and thickness scan by Dektak-type profilometer with 1 μm
conospherical t ip on C1A1/5wt% MP1100. Table 3.9 contains a list ing of surface
roughnesses of various tradename coatings marketed for tribological performace.
Fig. 3.19. (Left) C1A1/5wt% TE5069AN deposited by HVLP on G2 Duraba r
substrate. (Right) C1A1/5wt% MP1100 deposited by HVLP on G2 Durabar
substrate. The blue tape is used as a negat ive to establish coating thickness.
72
Fig. 3.20. Example roughness and coat ing thickness scans for batch of C1A1 / 5wt%
MP1100 coatings deposited from a 0.30g/mL solut ion on roughened G2 Durabar
substrate. Roughness step established by use of tape layer as negat ive.
Table 3.9. Tradename coatings marketed for tribological propert ies, their
abbreviated names used here and in some prior literature on coating tribological
performace and their Rq surface roughnesses.
73
Coating thickness was also established via scanning electron microscopy (SEM) as
shown in Figure 3.21. SEM imaging was generally conducted by Hitachi S4700
unless otherwise noted. The cross-sect ion image in Figure 3.21 is also notable in
that the cross-sect ioning was performed by band-saw and as can be seen, there was
no delaminat ion at the substrate-coating interface, which indicates good adhesion to
the substrate.
Fig. 3.21. SEM cross-sect ion image of C1A1/5wt% MP1100 ATSP coating on G2
Durabar substrate. Coating thickness of 20 -25 μm is demonstrated. Hitachi S4700
with accelerat ion voltage of 3 kV and working distance of 10 mm was used at a
magnificat ion of 450X.
The coating process demonstrated here was very flexible in terms of oligomer
structure that could be used as feedstock for the coating. All of the original Frich
and Economy [4] oligomers that have been examined to date are soluble in NMP to
74
at least 0.15 g/mL. It appears at least that below 2000 g/mol molecular weights this
rule holds
Fig. 3.22. Solvent-deposited cured ATSP coatings under cross-polarized transmitted
light. (Top Left) C1A1, (Top Right) C1A2), (Bottom Left) C2A1, (Bottom Right)
C2A2. All samples were 20 μm in thickness.
almost regardless of monomer unit including for oligomers bearing PFSA as a
backbone unit. Oligomer structure has had substant ial impact on coating
microstructure and, as will be shown later, both coating micro - and macro-
tribological properties. Figure 3.22 demonstrates the shift from amorphous to
liquid-crystalline oligomer-based coat ings and the apparent preservat ion of the
75
birefringence from into the cured structures. C1A1 demonstrates a completely
amorphous structure while C1A2 and C2A1 show a phase segregat ion between
amorphous and ordered regions. C2A2 possesses a fully birefringent morphology.
Specific ident ificat ion of ordered phase type may yield addit ional insight.
A C2A2 film was cast onto a glass slide and then the gla ss slide broken such that
the film broke in tension at room temperature. This was examined via SEM as
shown in Figure 3.23. Interest ingly, this demonstrated substant ial fibrillat ion at
some domain orientat ions and britt le fracture at others. C2A2 sample shown in
Figure 3.22 has a lower thickness than that of 3.23 and so the domains of 3.23 have
a relaxed growth constraint and can grow to a larger length scale.
In addit ion to other experimental evidence towards ITR as a chemical process
relevant to reordering and adhesive bonding, an interest ing phenomena was
observed (as shown in Figure 3.24) when a 50 μm C2A1 coating was observed on a
glass slide after having been cured to 270°C for 30 minutes in a convect ion oven –
which at this cure condit ion is already a crosslinked solid well below its glass
transit ion temperature and which at no temperature evidences melt ing. After
applying another 3 hours of cure at 330°C, we see an ent irely different morphology
within the same sample, from an “oil-in-water” phase segregat ion to a far more
“spider web” morphology. This change in microstructure appears to be an ent irely
solid-state phenomena that proceeds via ITR.
76
Fig. 3.23. (Top Left, Top Right, Bottom Left): SEM images of fibrillized region of
C2A2 coat ing under increasing magnificat ion. (Bottom Right) Surface of
commercially-produced aromatic polyester Vectran fiber exhibit ing similar
morphology [2].
77
Fig. 3.24. The same 50 μm thick neat C2A1 film after (Top) 270°C, 30 min cure and
(Bottom) an addit ional 330°C, 3 hour cure cycle.
78
3.8: Development of an electrostatic powder coating scheme for ATSP
To obviate the need for NMP, another coating pathway was examined. An init ial
trial was conducted wherein uncured CB and AB oligomer powders ground via a
Col-Int Tech Laboratory Grinder and then sieved through a 90 μm screen via shaker
table were blended at a 1:1 mass rat io and simply smeared across a roughened
stainless steel substrate. The samples were placed into a convect ion oven heated to
250°C for 10 minutes and then removed. The oligomers had obviously melted and
were adherent to the roughened substrate and passed the finger -nail test and
otherwise indicated that they had acceptable adhesion.
Having passed that init ial screening, a Gema bottom-fed electrostatic powder spray
gun was used to produce a stream of posit ively-charged powder which was then
picked up by a negat ively-charged conduct ive plate. A 200V charge was ut ilized to
build the uncured coating layer in all subsequent experiments. Here again CBA B
powder at a 1:1 rat io, ground, and sieved through a 90 μm screen (representative
particle size distribut ion shown below obtained by image analysis of transmitted
light micrographs of dispersed powder).
79
Fig. 3.25. Representat ive part icle size distribut ion of ground CBAB powder.
Subsequent to electrostatic powder deposit ion, panels shown in Figure 3. 26 were
hung from a hook inside a convect ion oven and allowed to melt at 230°C. Result ing
coating thicknesses as measured by magnet ic coating thickness gaug e (Reed
Instruments CM-8822) are shown in Figure 3.27. It can be noted that the
electrostatic powder coating experiment produced a glossy surface that easily
passed the fingernail test on all panels, including unroughened panels. Figure 3. 27
shows the cure cycle employed on all electrostatic-powder deposited coat ings
except where noted otherwise. For powder -coated samples, the cure cycle shown on
Figure 3.28 in a Thermo-Fisher convect ion oven was used except where specified
otherwise.
80
Fig. 3.26. (Top) Coated panels (2”x5”) after oligomer was allowed to melt for 10
minutes at 230°C. (Bottom) Curing condit ions, where 270°C was held for 30
minutes and 330°C for 3 hours.
81
Fig. 3.27. Coat ing thickness for samples shown in Figure 3.27 as determined by
magnet ic coat ing thickness gauge and profilometer .
Fig. 3.28. Representat ive trace of cure cycle used for electrostat ic ATSP powder
coatings.
82
As can be seen in Figure 3.27, there was approximately an 8% loss in thickness
upon cure for the init ial CBAB coated specimens, which corresponds closely to the
mass loss percentage upon cure. Similar to the solvent -borne coating deposit ion
system described above, all of the oligomers attempted proved usable with this
scheme. Prior to cure, all of the oligomers have very weak mechanical properties
and are easily ground. As well, all of the oligomers enter an obvious molten state by
230°C although generally the melt ing point appeared quite diffuse. Lubricat ing
adddit ives employed above were also generally usable however, a decrease in
concentration turned out to be necessary. Figure 3. 29 through 3.33 shows several
cure experiments of ATSP oligomeric powder observed by reflected light hot stage
experiment after electrostatic powder deposit ion onto aluminum foil. Samples were
ramped to indicated temperatures where they were held for 5 minutes before
proceeding to the next stage. It can be observed that there is a dist inct morphology
difference that evolves for the specimens that contain >4wt% of the MP1100 PTFE
addit ive. This morphology was generally referred to as the “loose” morphology as it
produces a rough surface texture. The loose morphology is created due to PTFE
having a relat ively high contact with the oligomer melt. As the oligomers melt, they
contact the PTFE. As the oligomer contacts the (relat ively fine part icle size) PTFE,
its flow and connect ion with other oligomer particles can be obstrcted due to the
high contact angle and high viscosity of the oligomers. This creates voids in the
microsctructure of the melt which get locked into place by the cure process. A
higher PTFE concentrat ion and a thicker layer of high PTFE concentration
83
C1A2/C2A1 tend to create the condit ions for the loose structure. CBAB tends
towards this set of circumstances at an even lower PT FE concentration due to the
higher flow temperature of its const ituent oligomers - hence why CBAB and
CB2AB2 were generally used at <2wt% rather than the higher concentration used
with the other oligomer systems. An example of the loose morphology can be seen
in Figure 3.31.
84
Fig. 3.29. Neat CBAB powder under reflected light hot stage experiment after 5
minutes at each temperature step. (Top Left) Deposited powder at RT. (Top Middle)
At 150°C. (Top Right) 230°C showing substant ial flow. (Bottom Left) 270°C.
(Bottom Right) 330°C.
Fig 3.30. Figure 3.31. CBAB/2wt% MP1100 powder under reflected light hot stage
experiment after 5 minutes at each temperature step. (Top Left) Deposited powder
at RT. (Top Middle) At 150°C. (Top Right) 230°C showing substant ial flow.
(Bottom Left) 270°C. (Bottom Right) 330°C.
85
Fig. 3.31. CBAB/5wt% MP1100 powder under reflected light hot stage experiment
after 5 minutes at each temperature step. (Top Left) Deposited powder at RT. (Top
Middle) At 150°C. (Top Right) 230°C showing substant ial flow. (Bottom Left)
270°C. (Bottom Right) 330°C.
Fig. 3.32. CBAB/5wt% MP1100 powder under reflected light hot stage experiment
after 5 minutes at each temperature step. (Top Left) Deposited powder at RT. ( Top
Middle) At 150°C there is obvious softening of the oligomer . (Top Right) 230°C
showing substant ial flow. (Bottom Left) 270°C. (Bottom Right) 330°C.
86
Fig. 3.33. Neat C1A1 powder under reflected light hot stage experiment after 5
minutes at each temperature step. (Top Left) Deposited powder at RT. (Top Right)
At 150°C there is obvious softening of the oligomer . (Bottom Left) 230°C showing
substant ial flow. (Bottom Right) Coating at 270°C fully transparent .
Powder-coated specimens of aluminum foil were cured to 270°C after which their
aluminum substrate was dissolved via 2N hydrochloric acid in deionized water.
Following this, isolated freestanding films were thoroughly rinsed in deionized
water followed by isopropyl alcohol and then dried at 70°C in air. Samples were
then examined via cross-polarized transmitted light microscopy as shown in Figure
3.34. Again as in the case for solvent -deposited coatings, the trend towards
increasing birefringence when using liquid crystalline, lower cross -link density
oligomers is preserved.
87
Fig. 3.34. Electrostatic powder-deposited cured ATSP coatings under cross-
polarized transmitted light. (Top Left) C1A1, (Top Right) C1A2), (Bottom Left)
C2A1, (Bottom Right) C2A2. All samples were 20 μm in thickness.
88
3.9: Macro-tribological Characterization of ATSP Coatings
Macro-tribological properties of ATSP-coated specimens (and therefore their
suitability as protective coatings in the complex condit ions experienced in the
compressor environment) were determined primarily through accelerated, high
pressure wear test ing via high pressure tribometer (HPT). The HPT features a
pressurized chamber which can be flooded with refrigerant gases as well as
thermocouple measurement of the near contact temperature (NCT). A list of
experimental parameters is featured in Table 3.10. HFO-1234yf is the trade name
for 2,3,3,3-tetrafluoropropene a potential “drop in” replacement for R-134A
(1,1,1,2-tetrafluoroethane) with a substant ially lower GHG equivalent number [122-
123]. Counterface was a 52100 hardened steel shoe, an actual compressor part used
in automotive swash-plate compressors [86].
Table 3.10. Condit ions for testing of solvent-deposited ATSP coatings.
89
Results generally demonstrate extraordinarily low cohesive wear of ATSP coatings
in comparison to state-of-art (Table 3.9) commercial coat ings marketed for
tribological properties. For unidirect ional motions, C2A1 w ith 5wt% MP1100 was
found to exhibit the lowest COF under both dry and lubricated condit ions, as shown
in Figure 3.35. Figure 3.36 demonstrates superior surface integrity of both C1A2
and C2A1 coat ings in unlubricated sliding condit ions at 6 MPa of contact pressure.
Moderate wear can be observed on the PTFE/MoS 2 surface and wear scratches
observed on the Fluorocarbon and PTFE/PEEK surfaces. The PTFE/Pyrrolidone
surface evidenced burnishing. Figure 3.37 shows the same coatings under the same
condit ions except now under boundary lubricat ion.
Fig. 3.35. COF versus sliding distance for various polymeric coat ings on G2
Durabar as measured by unidirect ional HPT with condit ions in Table 3.10.
90
Fig. 3.36. Optical microscopic images of unlubricated coating surfaces subsequent
to HPT wear experiments
Fig. 3.37. Optical microscopic images of boundary-lubricated coating surfaces
subsequent to unidirect ional HPT wear experiments
Again, ATSP coat ings exhibited a higher resistance to wear: no significant wear
mark was observed. Fluorocarbon and PTFE/PEEK coatings demonstrated a worse
performance than in the dry condit ion, which is thought to originate in the
migrat ion of the wear debris film under lubricated condit ions. PTFE/Pyrrolidone
91
coating evidenced an almost ident ical performance under both dry and lubricated
condit ions. Wear depth tracks for both condit ions of unidirect ional wear, as
observed by profilometry, are shown in Figure 3.38.
Fig. 3.38. Profilometry of tribological coat ings after HPT wear experime nt.
92
Fig. 3.39. COF versus sliding t ime for various polymeric coat ings on G2 Durabar as
measured by oscillatory HPT with condit ions in Table 3.10.
A significant reduct ion in frict ion (both in average and standard deviat ion) was
observed for the ATSP-based coatings. This suggests a synergist ic effect between
the lubricant and coating in these cases. PTFE/Pyrrolidone evidenced almost
ident ical performance between unlubricated and boundary lubricated condit ions,
indicat ing no real benefit from the presence of lubricant. The COF increased for
PTFE/MoS2 and PTFE/PEEK coatings again likely due to wear debris films
migrat ion via the lubricant. No surface damage was observed for the ATSP films in
the either lubricated or unlubricated condit ions. Moderate wear was observed on all
other coatings.
93
Fig. 3.40. Optical microscopic images of boundary-lubricated coating surfaces
subsequent to oscillatory HPT wear experiments
0 2 4 6 8 10-20
-16
-12
-8
-4
0
4
8
Wear
dep
th (
m)
Wear scan (mm)
Unlubricated
Boundary lubricated (PAG)
0 2 4 6 8 10-20
-16
-12
-8
-4
0
4
8
Wear
dep
th (
m)
Wear scan (mm)
Unlubricated
Boundary lubricated (PAG)
0 2 4 6 8 10-20
-16
-12
-8
-4
0
4
8
Wear
dep
th (
m)
Wear scan (mm)
Unlubricated
Boundary lubricated (PAG)
0 2 4 6 8 10-20
-16
-12
-8
-4
0
4
8
Wear
dep
th (
m)
Wear scan (mm)
Unlubricated
Boundary lubricated (PAG)
0 2 4 6 8 10-20
-16
-12
-8
-4
0
4
8
We
ar
de
pth
(
m)
Wear scan (mm)
Unlubricated
Boundary lubricated (PAG)
PTFE/C1A2
PTFE/C2A1
PTFE/MoS2 Flurocarbon
PTFE/Pyrrolidone
Fig. 3.41. Profilometry of tribological coat ings after oscillatory HPT wear
experiment.
94
Figure 3.42 summarizes the COF and wear results (as determined by integrat ion of
the volume of the wear tracks with t ime and load) of all coat ings undergoing HPT
testing and shows that ATSP-based coat ings occupy an excellent posit ion with
respect to both parameters.
Fig. 3.42. Summary of ATSP and assorted tradename polymeric tribological
coatings.
95
As part of the origins of the superior performance of the ATSP -based coat ings,
coatings at various stages of testing were examined. The phase segregat ion
observed under cross-polarized transmitted light manifested as surface segregat ion
with the more ordered and birefringent phase serving as the globular surface
portrusions (Figure 3.43). After a relat ively short sliding distance, the segregated
peaks become sheared off and the material amalgamates and init iates format ion of a
layer on the polymeric surface. After cont inuous sliding, this results in a deformed
and stable film layer on the polymer. This same behavior is seen on the counterface
(Figure 3.44), where a non-fluorinated stable layer forms part of the counterface
debris film which in this case forms a very stable and consolidated film. This
suggests that ITR was involved in this mechanism as near contact temperatures
were frequent ly observed to be up to 200°C throughout these experiments. As well,
blue-ing of the steel shoe provides a qualitat ive indicat ion of actual contact
temperature (NCT will tend to underest imate this number).
Fig. 3.43. Morphological analysis of solvent -deposited C1A2 + 5wt% MP1100
coating after unidirect ional wear experiments.
96
Fig. 3.44. (Top Left) SEM back-scatter imaging of 52100 steel counterface surface.
(Top Right) SEM image of 52100 steel showing topography. (Bottom) electron
dispersive spectroscopic scan of debris film indicated presence of CHO polymer
(ATSP) present and forming a stable film layer.
A similar study was undertaken on electrostatic powder coatings of ATSP. Contact
geometry was changed to a narrower pin such that contact pressure increased to
15.6 MPa and the pin metal was changed to 4340 steel. A similar process was used
97
as above for unidirect ional wear, results are summarized in Fig. 3. 45. All specimens
aside from C1FA2F and C2FA1F were used with MP1100 as lubricat ing phase.
Fig. 3.45. Summary of powder-deposited ATSP and assorted tradename polymeric
tribological coat ings. Black labels are for unlubricated and red labels are for
lubricated condit ions.
An extended durat ion unidirect ional HPT experiment was undertaken with CB2AB2
oligomer coated on a C932 bronze substrate for condit ions intended to replicate
submersible pumps utilized in the offshore energy-extract ion industry. An
extraordinarily high Pressure x Velocity-stability was observed for this system with
98
an incredibly low and stabilizing wear depth, which translates to a remarkable low
wear rate. Results and high level of surface integrity after these incredibly severe
and extended experiments are show in Figure 3.46. Table 3.11 summarizes these
results in terms of PxV values.
Fig. 3.46. Extended unidirect ional HPT test on CB2AB2 + 1.4wt% MP1100 on
C932 Bronze
Table 3.11. Summary of results on extended duration unidirect ional test ing.
99
3.10: Conclusions
ATSP solvent-borne and electrostatic powder coatings offer performance superior to
state-of-art tradename coat ings. ATSP electrostatic powder coat ings are now (with a
Tg of up to 307°C (compared to PEEK at 143°C) are now the highest temperature
electrostatic powder coatings seen in literature or available commercially by more
than 150°C, which offers new ranges of ut ility and performance to diverse
mechanical components.
100
CHAPTER 4
MICROMECHANICAL PROPERTIES AND ELASTIC RECOVERY OF
AROMATIC THERMOSETTING COPOLYESTERS
4.1: Introduction
Using a combination of large force range (~ 300 mN) and large displacement range (~ 7 μm)
indentation, the complete coating response could be measured together with the effect of the hard
substrate. The microstructural differences between imposed by changes in the oligomer
architecture were used to explain differences in the micromechanical properties of ATSP
coatings. The measured micromechanical properties, which could be readily obtained by
performing simple laboratory indentation tests, were successfully correlated with compressor
specific tribological experiments.
Seven different commercial polymeric coatings, namely, PTFE/Pyrrolidone-1 (DuPontTM Teflon®
958-303), PTFE/Pyrrolidone-2 (DuPontTM Teflon® 958-414), Resin/PTFE/MoS2 (Whitford
Xylan®
1052), PEEK/PTFE (1704 PEEK/PTFE®), PEEK/Ceramic (1707 PEEK/Ceramic
®),
Fluorocarbon (Impreglon® 218), and PTFE/MoS2 (Fluorolon
® 325) were used in this chapter.
Coatings blended with MoS2 can normally sustain high normal loads and may offer, at the same
time, low shear strength due to lamellar structure of MoS2, thus resulting in a lubricious low
friction surface. Several coatings based on ATSP were also selected for this comparative study.
These coatings were blended with Zonyl-type PTFE additives and MoS2 to implement a
lubricating phase. ATSP was also varied from a low to a high crosslink density by adjustment of
feed ratio of trimesic acid from C1A1 to C2A2.
101
Sample Oligomer mass feed ratio (C#/A#) Oligomer molar feed ratio (C#/A#) COOH / AcO xn (M̄c)N
ID# (g)
C1A1 1.105 1.000 1.00 0.1440 834.10
C1A2 1.140 1.000 1.33 0.1051 1142.82
C1A2+ 1.243 1.087 1.45 0.1063 1129.92
C1A2- 1.029 0.900 1.20 0.1030 1166.12
C1A2= 0.857 0.750 1.00 0.1010 1189.21
C2A1 1.080 1.000 0.75 0.1100 1117.30
C2A1- 1.224 1.133 0.85 0.1075 1117.30
C2A1= 1.440 1.333 1.00 0.1040 1091.91
C2A1+ 1.584 1.467 1.10 0.1020 1177.55
C2A2 1.117 1.000 1.00 0.0718 1672.84
Fig. 4.1 Variation of Mc with oligomer choice. Branch parameters as described in Chapter 2.
To better understand the overall tribological behavior, an attempt was made to investigate how
the frictional force is determined for sliding wear of a polymer surface against a metallic
counterpart. From the well-known model by Briscoe [81], the surface layer of the polymer
involved in the frictional process can be classified into two zones: the interfacial zone with a
depth of about 100 nm and the cohesive (subsurface) zone corresponding to the depth of the
coating thickness. Therefore, the frictional force resulting from the adhesion equals the product
of the “real contact area” of the interfacial zone (with the counterpart) and the “shear stress” of
the subsurface zone. This is assuming that the counterface is sufficiently hard in comparison to
102
the polymer-mating surface and undergoes only mild or no elastic deformation. Figure 4.2
shows a schematic diagram of the two-term friction model of polymeric surface.
Fig. 4.2. Briscoe’s two-term model of polymer interface showing interfacial and cohesive zones [81].
The frictional work at the interface is the result of adhesive interactions between the polymer
surfaces and the rigid asperities which depend on factors such as: the hardness of the polymer,
molecular structure, glass transition temperature, crystallinity of the polymer, surface roughness
of the counterface and chemical/electrostatic interactions between the counterface and the
polymer. For example, a soft elastomeric solid with a glass transition temperature well below
room temperature, would have a large adhesive component leading to increased friction. Beyond
interfacial work is the contribution of the cohesive term, which is a result of the plowing actions
of the asperities of the harder counterface into the polymer. The energy required for the plowing
action will depend primarily on the tensile strength and the elongation before fracture (or
toughness) of the polymer and the geometric parameters (height and the cutting angle) of the
asperities on the counterface [81]. In a normal sliding experiment, however, it is hard to
103
completely decouple the two terms (interfacial and cohesive) and therefore most of the data
available in the literatures generally include a combined effect.
A critical component of indentation measurements is the calibration of the tip, known as area
function. The classic tip area function routine, known as the Oliver-Pharr method [88], is a
compliance method wherein the mechanical properties are determined based on the previously
observed contact area of the probe tip to the sample, under a certain load. Since indentation
measurements give us only contact depth information, the tip area function correlating the
contact area and the contact depth needs to be determined/ known for the tip used in the
measurements. Note that contact depth hc is defined as the vertical distance along which contact
is made by the tip during loading, and is less than the maximum penetration depth (hmax) due to
the elastic property of the indented surface [88]. If the tip was manufactured to be perfect
without any defects (which is not the case), the area function will be simply the geometrical
shape function of a pre-specified tip such as: Ac(hc) = 24.5 hc2
in the case of the Berkovich tip.
However, due to tip imperfections, the area function usually takes the following polynomial
form. To determine the coefficients of the above equation, indentations at varying penetration
depths (corresponding to similar depth range as the desired measurements) are performed on a
standard material. Then, since the modulus of the standard material is known, the contact area
corresponding to each contact depth (Ac, hc) can be calculated and plotted as in Figure 4.4, and
finally, the coefficients are determined by polynomial curve-fitting.
Two important elements that affect the tip calibration and thus the material property
values are 1) the choice of the standard sample and 2) the contact depth range of calibration. In
this study, a modified bismaleimide polymer (manufactured by BASF Corp) with a reduced
104
modulus value of 5.0 GPa was used as the calibration standard, as its value is similar to the
polymeric coatings to be tested in this work. Note that we did not use the widely used fused
Quartz standard because its reduced modulus is significantly higher than our interest, namely
69.6 GPa. Calibration was performed with 22 varying loads from 1 mN to 800 mN (under the
exact same loading conditions as the measurements detailed later) resulting in 0.5 to 6.5 μm of
contact depths (Figure 4.3) which were of similar range to those of the actual measurements.
Once the tip is completely calibrated with the coefficients seen in Figure 4.3, this tip area
function is directly used for the calculation of the mechanical properties during the indentation
measurements. The detailed calculation process using the compliance method which involves
the determination of the elastic modulus and hardness exclusively from the initial unloading
portion of the load-displacement curve can be found in the literatures [88,140].
Fig. 4.3. Polynomial curve-fitting of measured contact area versus contact depth to determine the
coefficients of tip area function. Screenshot of Hysitron Triboscan analysis software.
105
A typical indentation experiment on a polymeric material involves engaging the probe tip to the
sample surface under a light load (2 μN), indenting to a pre-specified maximum load at a
constant loading rate (10 mN/s), holding at the peak load (5 sec in this case) to reduce creep
effects [81], and then withdrawing the tip at the same rate as during loading (~10 mN/s). This is
called trapezoidal loading profile, compared to the most commonly used triangular profile which
does not have a hold-time at the peak load. The initial set of indentations were performed on
PTFE/Pyrrolidone-2 coatings using both trapezoidal and triangular loading profiles with a
maximum load of 70 mN to examine possible creep behavior. For the triangular loading, three
different loading/unloading rates (2, 10 and 30 mN/s) were used to determine their effect on the
slope of the initial part of the unloading curve (and thus the calculated property values). For the
main indentation measurements: a trapezoidal loading profile at a constant loading/unloading
rate of ±10 mN/s was used for about 10 varying maximum loads between 5 to 400 mN (which
correspond within the calibration range). For each maximum load, six indentations were
repeated at different areas (resulting in approximately 60 indentations for each coating sample),
thus resulting in 420 indentations in total. Although this method of using single
loading/unloading indentations is time consuming, compared to the partial unloading method, it
is preferred for precise examination of polymer material properties. Also, in this work, single
indentations showed valuable phenomena for each coating, which correlated their structural
properties with their tribological behavior.
Figure 4.4 depicts typical ramp-loading scratch experiments that were used for characterizing the
coating’s. Pre- and post-scratch scans of the surface topography were also performed due to the
106
viscoelastic behavior and the relatively high hardness of the coatings. Therefore, each scratch
experiment consists of three steps: (1) a pre-scratch scan to measure the topography of the
original surface to be scratched, (2) an actual ramp loading scratch, and (3) a post-scratch scan to
measure the residual deformation of the scratched surface. All three steps were performed at the
exact same location. At the beginning and end of each step, a pre- and a post-profile scan, which
is about 12 % of the scratch length, is performed to correct the alignment of the data, as shown in
Figure 4.4(a). After the completion of a scratch test, the scratch profile (red line) and the
residual profile (green line) are normalized based on the reference original surface (black line) to
precisely quantify the true depth of scratch as well as the elastic and plastic deformations of each
coating surface, as seen in Figure 4.4(b). Along with the displacement plots, the in-situ lateral
and normal forces are recorded, thus resulting in the in-situ friction coefficient of the scratched
surface.
-200 -100 0-6
-4
-2
0
2
Depth
(m
)
Scratch Distance (m)
Pre-Scan
Scratch
Post-Scan
Pre-profile scan
Post-profile
scan
-200 -100 0-6
-4
-2
0
2
Depth
(m
)
Scratch Distance (m)
Scratch
Post-Scan
(a) (b)
Scratch direction
Fig. 4.4 Representative ramp-loading microscratch curves showing (a) low load pre-scratch scan (black),
in-situ scratch displacement (red), and low load post-scratch scan (green); (b) normalized scratch and
post-scratch scans.
107
A constant normal load of 200 µN was applied for both the pre- and post-scratch scans.
Each sample was scratched 4 times for each maximum load (three different maximum loads of
10, 20, and 30 mN, thus resulting in 12 scratches for each sample) to ensure repeatability and
establish the scatter in the results. The tip used for the scratch tests was a 60° conical tip with a
tip radius of 4.3 μm as seen in Figure 4.5.
Fig. 4.5 (a) Scanning electron microscopy (SEM) image of conical tip used for scratch testing in this
work, (b) zoomed in image showing the tip radius.
Analysis was performed on two representative PTFE- (PTFE/Pyrrolidone-2) and PEEK-
based (PEEK/PTFE) coatings to examine the main differences in their micromechanical
properties and correlation with their microstructure, as well as correlation with their tribological
performance (section 4.3.3). Figure 4.6(a) and 4.6(b) show the load-displacement curves of
PTFE/Pyrrolidone-2 and PEEK/PTFE coatings, respectively. The elastic modulus and hardness
values determined from these curves (six measurements for each maximum load instead of the
two shown in the figure) are plotted in Figure 4.6(c) and 4.6(d), respectively. Clear differences
on load-displacement curves between the two coatings could be observed; PTFE/Pyrrolidone-2
in Figure 4.6(a) exhibited repeatable load-displacement behavior, namely, the two curves at the
same maximum load showed exactly the same path. Also, note that the loading curves in this
108
case are almost identical, for all peak load experiments. Considering that each indentation was
performed at a different area of the coating surface, this repeatability implies very uniform
structure and mechanical properties over the whole coating surface area. At the same time, the
smoothly increasing loading curve shows their structural uniformity through the coating
thickness. These characteristics explain the small x- and y-axis error bars of the elastic modulus
and hardness of PTFE/Pyrrodlidone-2 coating, as seen in Figure 4.6(c) and 4.6(d). Thus,
PTFE/Pyrrolidone-2 coating can be thought to have amorphous-like microstructure.
In the case of PEEK/PTFE coating, the load-displacement curves showed completely
different paths for all indentations, and the slopes of the loading curves differ significantly. This
implies that the mechanical properties of PEEK/PTFE coating differ from location-to-location on
the surface, which can also be seen by the large error bars of elastic modulus and hardness of
PEEK/PTFE coating in Figure 4.6(c) and 4.6(d). Also, examining the loading curve, its slope is
not smoothly increasing, but changing at some penetration depth, which means its microstructure
is not uniform through the coating thickness. As the tip is penetrating the coating, it finds a
nonhomogeneous material, with different mechanical properties (either amorphous or
crystalline), showing the semi-crystalline microstructure of PEEK/PTFE coatings.
109
0 2 4 60
2
4
6
8
hc (m)
Er (G
Pa)
PTFE/Pyrrolidone-2
PEEK/PTFE
0 2 4 60
20
40
60
80
100
H
(M
Pa)
hc (m)
PTFE/Pyrrolidone-2
PEEK/PTFE
0 2 4 60
100
200
Load (
mN
)
Displacement (m)0 2 4 6
0
100
200
Load (
mN
)
Displacement (m)
(a) PTFE/Pyrrolidone-2 (b) PEEK/PTFE
(c) (d)
Figure 4.6 Load-displacement curves of (a) PTFE/Pyrrolidone-2 and (b) PEEK/PTFE coatings showing
their load-bearing properties. Extracted (c) reduced modulus and (d) hardness values.
Figure 4.7 Schematic of load-displacement curves with hold (a) and regraphed as a function of time (b).
110
0 2 4 60
2
4
6
8
hc (m)
Er (G
Pa)
PTFE/Pyrrolidone-1
PTFE/Pyrrolidone-2
PTFE/MoS2
Fluorocarbon
PEEK/PTFE
0 2 4 60
20
40
60
80
100
hc (m)
H (
MP
a)
PTFE/Pyrrolidone-1
PTFE/Pyrrolidone-2
PTFE/MoS2
Fluorocarbon
PEEK/PTFE
(a) (b)
Figure 4.8 Extracted (a) elastic modulus and (b) hardness values with respect to contact depth.
Figure 4.8 shows the elastic modulus and hardness values of 5 different coatings as a
function of contact depth. Each data point shows an averaged elastic modulus (y-axis) and
contact depth (x-axis) along with their standard deviations (error bars in both axes) from six
indentation measurements at a given maximum load. In all cases except PEEK/PTFE, the elastic
modulus exhibits constant values at lower contact depths around 1.5 to 2 μm, and then increases
with increasing contact depth due to the substrate effect as above. In the case of PEEK/PTFE
coating, due to its porous structure, the elastic modulus is rarely affected by the substrate, thus
resulting in constant values with increasing contact depth. Interestingly, it can be observed that
the elastic modulus of PTFE/Pyrrolidone-1 coating is significantly higher than that of other
coatings as seen in Figure 4.8(a) (note that values at higher than 1 μm contact depths were not
included in the figure due to excessive pile-up during indentation, and thus inaccurate values).
Note that excessive pile-up phenomenon was only observed with PTFE/Pyrrolidone-1 as seen in
Figure 4.13. The underestimated contact area due to the pile-up effects was clearly observed for
PTFE/Pyrrolidone-1 coating in Figure 4.9(a), compared to other coatings with no pile-up, for
111
example, PTFE/Pyrrolidone-2 shown in Figure 4.9(b). Usually, for polymer materials, a blunt
tip with around 20 μm nominal radius is recommended to simply compress the material under the
tip to avoid any plastic deformation. However, when sharp indenters like the Berkovich are
employed, indentation becomes both elastic and plastic, thus resulting in such plastic
deformation as pile-up or sink-in. In the case of pile-up where material plastically uplifts around
the contact impression, the actual contact area is larger than that predicted by the area function
discussed above. Therefore, both elastic modulus and hardness values are overestimated as can
be seen in Figure 4.8 for PTFE/Pyrrolidone-1 [ref]. Finite element simulations for conical
indenters (Bolshakov and Pharr, 1998) have shown that the ratio of final indentation depth, hf
(the depth of the indentation impression after unloading) to the depth of the indentation at peak
load, hmax can be used as an indication of when pile-up is an important factor. Pile-up is
significant only when hf /hmax > 0.7 and the material does not appreciably work harden. For such
materials, failure to account for the pile-up can lead to an underestimation of the contact area
deduced from indentation load-displacement data by as much as 60%, thus resulting in an
overestimation of the hardness and elastic modulus. When hf /hmax < 0.7, or in all materials that
moderately work harden, pile-up is not a significant factor, and the Oliver–Pharr model can be
expected to give reasonable results. In this work in all coatings, except PTFE/Pyrrolidone-1, the
latter is the case and thus pile up effects can be safely ignored.
112
0 100 200 300-20
-10
0
10
20
Scan Length (m)
Pro
file
(m
)
PTFE/Pyrrolidone-2
100 μm100 μm
(a) (b)Pile-up
3 μm 5 μm
0 100 200 300-20
-10
0
10
20
Pro
file
(m
)
Scan Length (m)
PTFE/Pyrrolidone-1
Figure 4.9 Cross-section line profile of indentation residual impressions: (a) PTFE/Pyrrolidone-1 and (b)
PTFE/Pyrrolidone-2 coatings.
Figure 4.10 ATSP Condition matrix and indentation images. Top left (a) C1A1 cured in vacuum for 6 hr
at 320°C. Top right (b) C2A2 cured in vacuum for 6 hr at 320°C. Bottom left (c) C1A1 cured in forced air
at 270°C for ½ hr. Bottom right (d) C2A2 cured in forced air at 270°C for ½ hr.
113
PTFE-, PEEK-, and ATSP-based polymeric coatings were subjected to microscratch testing to
investigate their viscoelastic (recovery) properties and to correlate these results with their
tribological behavior. Figure 4.11(a,c,e) and Figure 4.11(b,d,f) show representative microscratch
results obtained with ramp loading up to 10mN for PTFE/Pyrrolidone-2 and PEEK/PTFE
coatings, respectively. As described in section 5.2.3, the in-situ scratch profiles were normalized
using the pre-scan to correct for the surface topography. Shown in the Figure 4.11(c) and
4.11(d) are the in-situ deformations with red color (that includes both elastic and plastic
deformation) as well as the post scan with green color that shows the permanent (or plastic)
deformation only (the difference being the elastic recovery). In the case of PEEK/PTFE, most
deformation was plastic as seen in Figure 4.11(d). Also, compared to the smooth scratch curves
of PTFE/Pyrrolidone-2, Figure 4.11(c), PEEK/PTFE showed significant fluctuations. This can
be attributed to both its rougher surface, and also to its non-uniform (semi-crystalline) structure
which was also examined in Chapter 4. Along with the displacement curves, the in-situ COF
corresponding to the displacement curve (red) was also measured as seen in Figure 4.11(e) and
4.11(f). A relatively stable COF was observed for PTFE/Pyrrolidone-2 coating, while
PEEK/PTFE coating showed some fluctuation due to a rougher surface and possibly its semi-
crystalline microstructure as well.
Figure 4.12(a,c,e) and Figure 4.12(b,d,f) show the microscratch results obtained with
ramp loading up to 20 mN for 1% ATSP/PTFE and 5% ATSP/MoS2 coatings, respectively.
Compared to PTFE- and PEEK-based coatings, these coatings showed relatively smoother
scratch curves and thus stable friction coefficient during the scratch. Also, it was observed that
114
most of deformation was elastic for ATSP-based coatings showing significant amount of
recovery of deformed surfaces.
-100 -50 0
-4
-2
0
2
Depth
(m
)
Scratch Distance (m)
Pre-Scan
Scratch
Post-Scan
-100 -50 0
-4
-2
0
2
Depth
(m
)
Scratch Distance (m)
Scratch
Post-Scan
-100 -50 0
-4
-2
0
2
Depth
(m
)Scratch Distance (m)
Pre-Scan
Scratch
Post-Scan
-100 -50 0
-4
-2
0
2
D
epth
(nm
)
Scratch Distance (m)
Scratch
Post-Scan
-100 -50 00.0
0.5
1.0
1.5
2.0
CO
F
Scratch Distance (m)
PTFE/Pyrrolidone-2
-100 -50 00.0
0.5
1.0
1.5
2.0
Scratch Distance (m)
CO
F
PEEK/PTFE
(a) (b)
(c) (d)
(e) (f)
PTFE/Pyrrolidone-2
PTFE/Pyrrolidone-2 PEEK/PTFE
PEEK/PTFE
Figure 4.11 Microscratch curves from ramp-loading up to 10 mN for (a) PTFE/Pyrrolidone-2 ((c) after
normalization), and (b) PEEK/PTFE ((d) after normalization). In-situ friction coefficient of (e)
PTFE/Pyrrolidone-2 and (f) PEEK/PTFE coatings monitored during the actual scratch cycle.
115
-200 -100 0-6
-4
-2
0
2
De
pth
(m
)
Scratch Distance (m)
Pre-Scan
Scratch
Post-Scan
-200 -100 0-6
-4
-2
0
2
De
pth
(m
)
Scratch Distance (m)
Scratch
Post-Scan
0 100 2000.0
0.5
1.0
1.5
2.0
CO
F
Scratch Distance (m)
1% ATSP/PTFE
-200 -100 00.0
0.5
1.0
1.5
2.0
CO
F
Scratch Distance (m)
5% ATSP/MoS2
0 100 200-8
-6
-4
-2
0
2
De
pth
(m
)
Scratch Distance (m)
Pre-Scan
Scratch
Post-Scan
0 100 200-6
-4
-2
0
2
De
pth
(m
)
Scratch Distance (m)
Scratch
Post-Scan
1% ATSP/PTFE 5% ATSP/MoS2
5% ATSP/MoS21% ATSP/PTFE
(a) (b)
(c) (d)
(e) (f)
Figure 4.12 Microscratch curves from ramp-loading up to 10 mN for (a) 1% ATSP/PTFE ((c) after
normalization), and (b) 5% ATSP/MoS2 ((d) after normalization). In-situ friction coefficient of (e) 1%
ATSP/PTFE and (f) 5% ATSP/MoS2 coatings monitored during the actual scratch cycle.
116
0 10 200
40
80
120
Ela
stic R
ecovery
(%
)
Normal Load (mN)
PTFE/Pyrrolidone-2
PEEK/PTFE
1% ATSP/PTFE
5% ATSP/MoS2
PTFE/
Pyrrolidone
PEEK/
PTFE
1%ATSP
/PTFE
5%ATSP
/MoS2
0
2
4
De
form
atio
n (m
)
Total
Plastic
PTFE/
Pyrrolidone
PEEK/
PTFE
1%ATSP
/PTFE
5%ATSP
/MoS2
0
2
4
De
form
atio
n (m
)
Total
Plastic
PTFE/
Pyrrolidone
PEEK/
PTFE
0
2
4
De
form
atio
n (m
)
Total
Plastic
1%ATSP
/PTFE
5%ATSP
/MoS2
PTFE/
Pyrrolidone
PEEK/
PTFE
0
2
4
De
form
atio
n (m
)
Total
Plastic
1%ATSP
/PTFE
5%ATSP
/MoS2
PTFE/
Pyrrolidone
PEEK/
PTFE
1%ATSP
/PTFE
5%ATSP
/MoS2
0
2
4
De
form
atio
n (m
)
Total
Plastic
PTFE/
Pyrrolidone
PEEK/
PTFE
1%ATSP
/PTFE
5%ATSP
/MoS2
0
2
4
De
form
atio
n (m
)
Total
Plastic
(a) 5mN (b) 10mN
(c) 15mN(d)
Figure 4.13 Total (penetration depth) and plastic (permanent) deformation of four different polymeric
coating surfaces at (a) 5mN, (b) 10mN, and (c) 15mN normal load; (d) elastic recovery of coatings with
respect to normal loads.
Triboindenter testing up to high loads (Figures 4.15 through 4.17) demonstrates that there exists for ATSP
a specific contact stress through which deformation is recovered in both indentation and scratch-type
experiments. This is compared in Table 1 for C1A1 with several other neat polymers and tradename
coatings a discrete event wherein a high level of elastic recovery switches to a plowing mode (illustrated
in Figure 4.19). Figure 4.18 shows an example from this table of a neat P84 polyimide coating verus
C1A1. This generally demonstrates that C1A1 has an incredibly high cohesive strength and will recover
117
deformation. This is thought to be in part due to the high cross-link density and the available microscopic
deformation mechanisms shown in Figure 1.1 around the ester bond creating a constrained conditioned
that requires a very high stress to produce rupture and a transition to a plow-deformation.
0 10 20 300.0
0.2
0.4
0.6
0.8
1.0
CO
F
Normal Load (mN)
PTFE/Pyrrolidone-2
PEEK/PTFE
1% ATSP/PTFE
5% ATSP/MoS2
Figure 4.14 In-situ friction coefficient vs. normal loads of 4 different polymeric coating surfaces
measured during actual scratch cycle.
118
Figure 4.15 Example nanoscratch data from Hysitron triboindenter on C1A1-type coating with 5wt%
PTFE. Top left is reflected light optical image at 20X. Top right is COF versus scan length. Bottom left is
normal displacement of indenter tip versus scan length. Bottom right is normal force applied by indenter
tip versus scan length. Specimen displays stead-state non-failure wear COF and sharp critical force.
119
Figure 4.16 Example nanoscratch data from Hysitron triboindenter on C2A2-type coating with 5wt%
PTFE. Top left is reflected light optical image at 20X. Top right is COF versus scan length. Bottom left is
normal displacement of indenter tip versus scan length. Bottom right is normal force applied by indenter
tip versus scan length. Specimen does not display a stead-state non-failure wear COF and evidences no
sharp critical force.
120
Figure 4.17 Scratch image summary for ATSP coatings.
Table 4.1. Condensed scratch parameters versus fully cured (under vacuum) C1A1.
121
Fig. 4.18. Example scratch for P84 versus C1A1 ATSP. Tip direction proceeds as arrow
indicates.
122
Fig. 4.19. Cartoon of transition from an elastic and recoverable mode to plowing deformation
mode in ATSP coatings.
Electrostatic and solvent-borne deposition of a variety of different ATSP formulations (C1A2,
C2A1, C1A1, C2A2 and CB2AB2 with 5% PTFE) were also examined. Figure 4.20 illustrates
the results of scratch experiments performed on ATSP solvent-borne and electrostatic spray-
coated panels. The COF results for electrospray coated panels are qualitatively in-line or even
better indicating that the electrostatic deposition technique can produce tribologically robust
coatings.
0 100 200 300 400
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
CO
F
Lateral Displacement (µm)
C1A1+5wt% PTFE
C1A2 +5wt% PTFE
C2A1 +5wt% PTFE
C2A2 +5wt% PTFE
CB2AB2+5wt% PTFE
0 100 200 300 400
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
CO
F
Lateral Displacement (µm)
E-C1A1 + 5wt% PTFE
E-C1A2 + 5wt% PTFE
E-C2A1 + 5wt% PTFE
E-C2A2 + 5wt% PTFE
E-CB2AB2 + 5wt% PTFE
Fig. 4.20. Scratch experiments results conducted on ATSP-coated Q-panels; (left) solvent-borne,
(right) electrostatic deposition.
123
CHAPTER 5
ADVANCED CHARACTERIZATION OF CONTINUOUS FIBER / ATSP COMPOSITES
5.1: Introduction and Background
As discussed in Chapter 1, the high mechanical properties of these materials are seen to arise from their
orientational self-ordering [1] and high rigidity of their constituent units. Examining the simplest of the
class of LCPs, poly(HBA) as shown in Figure 1.1, the phenylene units are prevented from rotating and
conformational freedom is only available through crankshaft motion about the ester bond [38].
In addition to their robust mechanical properties (see Chapter 1), LCPs are also known to have excellent
chemical resistance, high heat deflection temperatures and flame resistance, and hydrolytic stability [1,
2,9,11]. These features make them attractive as matrices in high performance composites. To date though,
no LCP is known to have been successfully marketed as a matrix for continuous fiber reinforced
composites.
A number of oligomeric systems with varying functionality and molecular weight have been developed
and characterized. The oligomers can be tailored to be liquid crystalline by adjusting the molar ratio, thus
providing unique advantages compared to conventional polymers. In addition, the number of functional
groups can be modified to affect the crosslink density. Varying the molecular weight adjusts the
mechanical properties of the neat resin and composite .
ATSP can be produced in either amorphous or liquid crystalline formulations by selection of appropriate
oligomers. Oligomers which do not display birefringence individually in their melt form amorphous
polymers when cured while those that do display birefringence produce polymers which are birefringent
after full curing[38,51]. Birefringent oligomers have similar molecular weights as isotropic oligomers but
with a lower feed ratio of the crosslinker (TMA) which results in a longer chain length between
124
crosslinks, as shown in Table 2.2. Figure 1.4 displays the difference in birefringence between amorphous
and liquid crystalline fully cured ATSP.
One aspect where the tendency of LCPs to locally match coefficients of thermal expansion to carbon
fibers (as discussed in Chapter 1) may find utility is in areas where a composite structure is exposed to a
high degree of thermal cycling. An example application of this is reusable composite cryogenic fuel
storage tanks. These are large filament wound structures (see Figure 5.1), while composite articles in this
study will not be filament wound, generally comparable properties can be obtained and resin-dominate
failure mechanics still hold [95,110].
Fig. 5.1. Composite filament wound cryogenic storage tank under fabricat ion for
NASA’s Space Launch System (SLS). From [131].
125
Local orientation of LCPs around carbon fibers has been observed in the literature via x-ray diffraction
(XRD) [18], as we all nuclear magnetic resonance (NMR) [38] and polarized light microscopy [17,51].
XRD Observations by Chung, et al observed that the degree of orientation is very similar for both the
LCP molecules and the carbon basal plane. They note that alignment parallel to the carbon surface is the
most stable state for liquid crystal domains. This then presents the question of whether there is any
difference in the mechanical bonding between fiber and matrix for matrices which are chemically similar
to the LCP but are amorphous in the bulk material. The objective of this study is to investigate whether
there is any significant difference in the strength of the bond between fiber and matrix as distinguished
between amorphous and liquid crystalline formulations of ATSP and whether a highly oriented fiber
surface (such as found on graphite fibers) is a strongly influencing factor. That is, do amorphous fibers (in
this case glass) evidence the same difference, if any, in the strength of the bond between fiber and matrix
when the ATSP composition is varied between amorphous (C1/A1) and liquid crystalline (C2/A2).
-0.64 ppm/K
Repeated Thermal cyclesΔT ≈ 270K
50-120 ppm/K
Fig. 5.2: Microcracks in an Epon 828/IM-7 carbon fiber composite produced by
thermal cycling.
126
As well, lightweight structural concepts are necessary to enhance and enable space transportation for
future exploration and science missions. Reduced mass of launch vehicles yields increased payload mass
fraction, which is a critical need for all in-space systems. In addition, resilient cryo-tank concepts will
also improve launch vehicle safety and potentially lead to orbital fuel storage tanks.
Advanced composite materials processable by cost-effective manufacturing play an important role in
developing lightweight structures for future space and planetary systems. With the growing demand for
improved performance, especially in aerospace, advances in polymer systems with extreme thermo-
mechanical properties are critical in providing: (1) excellent retention of mechanical properties in high
temperature environments and (2) high resistance to microcracking at cryogenic temperatures for
advanced NASA requirements as described in NASA’s Roadmap Technology Area 12 [133].
Development of material technologies for lightweight structures is critical for the aircraft industry to
facilitate next generation aerospace missions. Although polymer resins like epoxies are used extensively,
there is still a need for improved high temperature systems that are stable in aggressive environments.
Design of thermally stable polymer matrices > 350 °C is an important key in production of structural
components for high-speed aircraft and lightweight components near the engine [12-13,110]. The most
widely used resins, epoxies, are stable only up to 100-120 °C for long-term usage [95]. The phenylethynyl
terminated imide oligomers developed at NASA Langley have a degradation mechanism which is evident
as low as 177 °C [134]. Additionally, the imidized oligomers employed in several NASA studies [13-17]
have high melt viscosities (~ 4,200 cP) compared to that needed for successful vacuum assisted resin
transfer molding (VARTM) that yields low porosity and high mechanical properties. Other polymeric
resin systems that have been designed for high temperature stability are phenyl based epoxies,
polyimides, bismaleimides, PEEK and Vectran. However, these polymeric systems generally possess
either limited thermal properties or have extremely high melt viscosities that preclude fabrication of low-
porosity, high quality composites [9]. Additionally, thermoplastics like PEEK provide an advantage over
commercially extant thermosets with improved fracture toughness but their Tg is limited to 150 °C.
127
Some applications, like future supersonic transportation, require composite materials that are not only
stable at high temperatures but also with negligible microcracking when cycled over a range of
temperatures (thermal fatigue resistance). Thus the design of next generation polymer composite fuel
tanks as well as structural composites for supersonic vehicles relies on thermally stable systems that are
also thermal fatigue resistant [1,10]. Fiber reinforced composites cured at elevated temperatures contain
thermal residual stresses after fabrication because of a coefficient of thermal expansion (CTE) mismatch
between the carbon fiber (negative CTE) and matrix (positive CTE), and a mismatch in the CTE between
plies with different orientations, both of which cause formation of microcracks. Microcracking can cause
degradation of mechanical properties such as strength and stiffness [11] and an increase in gas
permeability [12].
5.2: Single Fiber Fragmentation
One of the most reliable microtest method available [132] that measures fiber-matrix interface strength is
single fiber fragmentation test (SFFT), of which a schematic is shown in Figure 5.3. In SFFT, a dogbone
of polymer bearing a single fiber at the centroid of the dogbone cross-section is strained to at least three
times the elongation at failure of fiber it bear. The fragment length of the fiber is then observed. The
fragment lengths (l), fiber tensile strength (σf), and fiber diameter (d) are then related to the interfacial
shear strength.
(1)
128
C1, C2, A1, and A2 oligomers (solids at ambient temperatures) were mechanically ground into a fine
powder and purified using a 3:1 methanol:water solution in a Soxhlet extractor whose primary bulb was
set by thermocouple-controlled heating mantle set at 70°C for 24 hours. The oligomers were then vacuum
dried at 70°C for 48 hours.
Oligomeric solutions in N-methyl-2-pyrrolidone (NMP) were prepared by first blending dried carboxylic
acid and acetoxy-end-capped oligomer powders at a 1.1:1 carboxylic:acetoxy to reflect the higher average
molecular weight of the carboxylic oligomers. Oligomers were then weighed out to 0.3 g and dissolved
into 10 mL of NMP stirred at 75°C.
Single filaments of Hextow IM-7 carbon from Hexcel and unsized Advantex fibers from Owens Corning
were separated from their tows and individually laid onto one of the specially machined PTFE molds.
Fig. 5.3: Schematic representation of single fiber fragmentation process.
From [132].
129
Great care was taken to only touch the ends of the separated fibers. Fibers were secured at the ends of the
mold in a crevice that extends to the midpoint of the mold cavity using modeling clay.
Oligomer solution was carefully transferred to fiber surface using a fine metal-tipped syringe. Solution
was observed to wet and coat the fibers within 15 seconds under a stereoscopic bench microscope. After
coating all 8 fibers per mold, molds were placed in a thermocouple-controlled oven programmed with the
temperature profile shown in Figure 5.9 under vacuum. Molds bearing ATSP-coated fibers were then
removed from the vacuum oven. Previous investigation by AFM had shown coatings produced in this
manner had a thickness of up to 600nm.
The molds were then filled with a degased 31:32:32:37 mixture of EPON 828, HELOXY 505, and
Lindride 6K. Care was taken to pour slowly and minimally disturb the fiber. Epoxy dogbones were then
cured at 120°C for 4 hours before removal by bending mold along the transverse direction.
Produced specimens were then examined for well-aligned fibers in the center of the dogbone. Specimens
which failed to meet this criteria were discarded.
The specimens were then tested in a microstraining apparatus capable of applying sufficient load to
fracture the coupons. Specimens were clamped at the ends and given an elongation of 20%. Fiber
fragment lengths were then observed and recorded using a transmitted light polarizing microscope with
one polarizer below and one polarizer above the test coupon to see birefringence.
Fiber fragment lengths (as in Figure 5.4) were digitally captured and analyzed using the pixel counter
function in NSF Image J. Figure 5.5 shows the histogram of C1/A1 on Advantex glass fibers. This data is
summarized in Table 5.1. Weibull analysis of the was performed in OriginPro 8.6 with the critical length
estimated by equation (2):
(2)
130
Where α is the scale factor and β the shape parameter computed via a gamma function to arrive at the
mean E[lc].
The IFSS of the fibers was then calculated using equation (1) and are likewise reported in Table 5.1.
Table 5.1. Summarized mechanical properties of amorphous and liquid crystalline ATSP coupling agent.
C1/A1 C2/A2*
Advantex
α = 0.49
β = 2.76
lc = 0.436 mm
τi = 63.9 MPa
α = 0.43
β = 3.11
lc = 0.384 mm
τi = 72.8 MPa
IM-7
α = 0.25
β = 3.55
lc= 0.225 mm
τi = 53.5 MPa
α = 0.16
β = 3.29
lc = 0.143 mm
τi = 84.2 MPa
IFSS for all ATSP coated fibers was very high compared to published values for other fiber/matrix
systems for all combinations of the ATSP fiber/matrix interface. Previous to this study, there does not
appear to be any published information for IFSS of LCP fiber/matrix interfaces. One study [136] of
Kevlar 49 fibers embedded in an epoxy matrix presents IFSS values ranging up to 35 MPa, while IFSS
values in the range of 20-30 MPa are fairly common for glass fiber and epoxy systems [137].
131
The primary trend observed in these values is that ATSP formulations that are liquid crystalline in the
bulk polymer exhibit significantly higher IFSS values over their amorphous counterparts. The physical
reasoning for this is that if the polymer at the surface has a higher degree of orientation (Q), as
characterized by equation (3) and observable via XRD enables a stronger self-reinforcing effect.
(3)
Fig. 5.4: Fiber fragment length by cross-polarized transmitted light microscopy. Arrows
indicate fiber breaks. Scale bar is 500 μm.
132
One trend that is not definitively displayed is a significantly higher interfacial shear strength for highly
oriented fibers (IM-7 carbon fibers) as compared to the amorphous Advantex glass fibers. This indicates
that a highly oriented fiber surface does not appear to be required to induce local ordering of the LCP.
The very high IFSS values indicate that in continuous fiber composites based on ATSP, failure should be
highly resin dominated rather than dominated by fiber debonding and pull-out. This is corroborated by
SEM images of ATSP/C-fiber which show high cohesion of the matrix to individual fibers and fracture
surfaces which indicate matrix dominated failure.
0.0000 0.0005 0.0010 0.0015 0.0020
0
20
40
Count
Length (m)
Figure 5.5: Histogram of C1/A1 on Advantex glass fiber fragment length.
133
In Figure 5.6, on the right of the 90° tow, the fracture occurred at a lower level than that on the left side.
This results in more of the 90° fibers being visible. The fibers in the 0° tow are mostly held together as a
single consolidated bundle with very few stragglers. This suggests that the bonding between the resin and
the fiber surface was strong and was able to resist fracturing during the testing. The strength of the bond is
likewise shown by the few fibers that have been slightly pulled out and still have resin adhered to them.
Figure 5.7 likewise shows resin dominated interlaminar failure. In this case, a crack runs between the 0°
and 90° plies in a resin-rich region. Figure 5.8 is a higher magnification view of Figure 5.6 showing good
wetting of resin even near the fiber fracture surfaces.
Fig. 5.6: Interface between 0° and 90° fiber tows showing river lines and
good adhesion in the 0° tow.
134
Fig. 5.7: Area where two plies meet, showing resin-rich areas between plies
and a crack between the 0° and 90° of the top ply.
Fig 5.8: Higher magnificat ion of 0° tow in Figure 9 with good wetting of
resin between fibers.
135
5.3: Characterization of ATSP as the Resin-phase for Composite Cryogenic Storage Tanks
ATSP oligomer prepregs was prepared following closely from Parkar, et al [16]. Equimolar
blends of ATSP oligomer powders was added to n-methylpyrrolidone (NMP) and stirred at 70
°C until a homogeneous solution is produced. It has previously been found [9,16,17] that
solutions of 0.5 grams of oligomer per mL of NMP will produce composites with a 55 vol% fiber
fraction when an excess of 20% oligomer mass (with a cured density of 1.35 g/cm3) to that
necessary from mass balance. Therefore, 2.22 mL of ATSP oligomer solution was infiltrated into
each gram of a unidirectional mat of IM-7 carbon fibers from Hexcel. Prepregs will then be
desiccated of NMP to a leathery state wherein the composite prepreg has excellent drape.
To process the composites and remove the volatiles, a custom enclosed Carver hot press was
used. Two 1.27 cm thick steel plates was used in order to protect the platens in the hot press. A
Kapton film and fiberglass reinforced Teflon peel-ply was placed on either side of the
composites to isolate it from the steel plates. The composite specimens will then be treated to
the temperature cycle shown in Figure 5.9. At 205 °C, the vacuum pump was engaged to remove
excess NMP. A pressure of 0.7 MPa was applied at a sample temperature of 330 °C. After the
cycle is complete the sample was allowed to cool in the hot press to room temperature and then
removed for characterization.
136
Fig. 5.9. Layup for ATSP composite cure (top, left), final thick cured ATSP composite
specimen (top, right), and cure cycle for ATSP composite (bottom)
Composite specimens produced by this route will then be joined to desired specimen thickness
by use of ITR to produce thick consolidated specimens. Individual plies were stacked together
and exposed to the heat and pressure cycle described above, after which they will have a
seamless bond with excellent Short Beam Shear Stress (SBS) (above 60 MPa) [9]. Plies were
arranged to produce [0/90]s composites of approximately 2 mm in thickness as in Moll, et al [18]
for comparative purposes.
VARTM was examined using the same unidirectional IM-7 carbon fibers and oligomer solutions
from Task 2. The process will follow prior reports on VARTM of other high temperature
polymer matrix composites [12-15] but modified to suit the chosen ATSP chemistry. Oligomer
solutions were heated to 70 °C to yield viscosities ranging from 100-400 cP. This value is stable
indefinitely as the cure reaction does not substantially advance at this temperature. The solution
will then be infiltrated into a fiber preform using a VARTM apparatus. A stainless steel tool
137
plate was used as the base of the VARTM setup with Kapton polyimide film used for the inner
and outer vacuum bag and Aluminum screen mesh as the flow media. High temperature sealant
putty was used to ensure a vacuum-tight seal between tool plate and bagging. Tooling with fiber
preforms (of [0/90]s layup) inside the polyimide bagging will then be transferred to a Thermo
Scientific convection oven maintained at 70 °C and attached to vacuum apparatus and the heated
container of oligomer solution. The valve connecting the apparatus to the vacuum pump was
then be opened to allow the resin solution to be forced in and wet out the fiber preform. Once the
resin infusion is complete, a cure cycle analogous to that shown in Figure 5.9 was initiated.
To determine the volume fraction of fibers, resin, and voids (following from ASTM D3171), a
sample representative of each chemistry and processing method will first be weighed and its
volume determined by an Archimedean procedure and then digested using concentrated sulfuric
acid heated to about 80 °C. The samples were left in the acid for sufficient time (from prior
experience 8-12 hours for ATSP resins) to remove all of the resin from the fibers. The solutions
will then be filtered to recover the fibers and rinsed with copious amounts of water before being
put in an oven to dry. The fibers will finally be weighed on a balance and the volume fraction of
fibers, resin, and voids was calculated and reported.
ATSP is a unique high temperature matrix which shows potential for use as a high temperature
stable matrix for carbon fiber composites for structural applications. Use of these composites for
structural applications requires a thorough validation of the mechanical properties of bulk ATSP
as well as ATSP/C composite lamina. Various test standards were used for characterizing this
continuous fiber composite and neat resin. Tests are designed to study the behavior of
composites in different thermal environments to provide insights for Phase II and other future
works.
138
Dynamic mechanical analysis (DMA) was performed in a TA Instruments DMA Q800 on neat
ATSP resins to obtain the storage modulus and Tg following ASTM standard D5023-07. After
conditioning at 23 °C and 50% relative humidity for a minimum of 48 hours, the samples (12.5
mm x 25 mm strips) were loaded in 3-point bending configuration with a span of 25 mm. The
cryogenic reservoir was filled with liquid nitrogen and the temperature of the sample was ramped
from -196 °C to 150 °C at a rate of 1 °C/min. The storage and loss moduli was recorded as a
function of temperature. The glass transition temperature (Tg) was reported based on the peak in
the tangent of the phase angle between room temperature and 300 °C. Any observation of a
cryogenic-range Tβ and Tγ will likewise be reported based on lower tangent peak positions.
Thermal stability of representative clippings of ATSP composite specimens was measured in a
TA Instruments TGA 2950 ramped at 10 °C/min to 600 °C in air, N2, and CO2 to replicate a
wide range of potential mission atmosphere conditions. The trace of the mass loss/temperature
derivative curve as well as the total mass loss was reported. Additional high temperature hold
experiments at 371 °C for one hour in air was conducted with mass loss as a percentage reported
as point of comparison with VARTM polyimide composites from Cano et al. [13-14].
Mechanical strength and modulus of the composites was determined by short beam shear
strength (SBS) following ASTM D2344, and flexural strength and modulus following ASTM
D790. An Instron mechanical testing machine with temperature shroud was used with a test
speed of 1.27 mm/min. Flexural tests will carried out at a test speed of 0.76 mm/min using the
same load cell and temperature shroud. Temperature points of -90 °C, RT, and 180 °C was used
for this matrix of experiments for comparison with state of art composite materials and published
literature [13-17].
139
Fracture toughness was determined by double cantilever beam (DCB) configuration specimens
with a pre-crack introduced during the cure process by thin polyimide film to conform to ASTM
D5528. An Instron machine operating in displacement-controlled mode with a crosshead
displacement speed of 0.67 mm/s was used with an end-tab configuration affixed by solid state
bonded ATSP applied to the end tab to provide a high temperature adhesive [11]. Composite
specimens was 20 to 25 mm wide and at least 125 mm long and 3 mm thick. Temperature points
of -90 °C, RT, and 100 °C was used for the matrix of experiments.
CTEs for composite specimens was measured on a TA Instruments 2940 thermomechanical
analyzer (TMA). Samples was sectioned into 10 mm square blocks and then ramped in the TMA
at the rate of 10 °C/min from RT to 300 °C. A force of 50 mN was maintained throughout the
test. CTE was examined in this fashion for both longitudinal and transverse sections of ATSP
composites. This was compared to state of art composite materials for aerospace and especially
cryogenic and high temperature applications.
Carbon fiber-reinforced composites used in aircraft engines and other space applications must be
stable and capable of functioning in high temperatures. The main concern is the thermal
decomposition and thermal oxidative degradations at their service temperature that will affect
performance. In addition, microcracking can be observed during changes in temperature and
repetitive thermocycling, which leads to degradation in mechanical properties and an increase in
composite permeability. Microcrack development forms a penetrating network through the
thickness of the composite leading to cryogens leakage through the tank wall and limiting their
use in cryotanks [138].
140
Several composite specimens from each resin formulation was subjected to 200 h at 280 °C in
air. Samples tested in this manner will then be subjected to short beam shear experiment.
The thermal fatigue characteristics of ATSP versus a reference high performance epoxy
composite currently used in prototype cryogenic tanks (provided by NASA) was studied by
thermal cycling over a range of temperatures. A schematic of the cycler is shown in Figure 5.10,
which consists of an oven and a cryogenic bath. First an ATSP composite sample was examined
under an optical microscope to verify that no cracks are present prior to thermal cycling. Then a
composite sample is placed in a wire basket and cycled repeatedly from oven to cryogenic bath.
A heated copper barrel can be used to heat up the sample. The temperature of the barrel is set
using a temperature controller and a thermocouple. The sample was heated up to the ambient
oven temperature (35 °C) in 8 minutes (based on prior experience), and then cooled to liquid
nitrogen temperature (-196 °C) again for another 8 minutes.
Fig. 5.10. Schematic of the thermal fatigue apparatus.
141
Sample permeability was evaluated in a pressure cell apparatus. In this experiment, nitrogen is
flowed into the cell at a preset pressure recorded by a transducer. If a network of cracks is
present in the sample, nitrogen gas will flow through the sample and results in an increased
output pressure on the opposite side of the sample. The output pressure is measured and recorded
by a second pressure transducer. The test was run for several minutes to monitor the
development of the output pressure evolution with time. First the manufactured samples was
tested for leaking then the process of cryogenic cycling and permeability testing was repeated
until the sample leaks. Before performing the leak test, samples was placed in a desiccator for
several hours to ensure all condensation is eliminated.
Current production of ATSP composites has been via a cured prepreg route, which are then
solid-state bonded into fully condensed laminates via ITR, all the mechanical characterization,
TGAs and volume fraction calculations reported here are for the ATSP/C composites made via
prepreg route. However for many larger structures with complex shapes, vacuum-assisted resin
transfer molding (VARTM) is preferable in terms of ease-of-fabrication. In this study, VARTM
process will be also examined.
Figure 5.11 shows the schematic diagram of fabricating ATSP carbon fiber composite. A lamina
of ATSP was made by first wetting the fibers (unidirectional fabric of Sikawrap 103C by
Hexcel) with a 50% w/v ATSP solution in NMP and then curing the fabric using the cure cycle
shown in Figure 5.9 in a custom-built vacuum-enclosed hot press. A Kapton film and fiberglass
reinforced Tefon peel-ply was used to isolate the composite from the steel plates (Figure 5.9).
The steel plates were brought together at the middle of step 2 (202 °C). A minimal amount of
pressure was applied to remove any residual solvent present. Pressure of 100 psi was applied at
the start of the last step (330 °C). The curing reaction was carried out in vacuum to facilitate
142
removal of the byproduct acetic acid and the solvent. Figure 5.13 shows a solvent impregnated
carbon fabric with ATSP resin before and after curing. The average thickness of the cured
composite lamina is 0.5 mm.
Fig. 5.11. Schematic diagram of fabricating ATSP carbon fiber composite.
Fig. 5.12. Enclosed hot press used for composite processing.
143
Composite specimens produced by this route are then joined to desired specimen thickness by
use of interchain transesterification reactions (ITR) to produce thick consolidated specimens.
Individual plies are stacked together and exposed to the heat and pressure cycle described above,
after which they will have a seamless bond (Figure 5.11). This was done by heating the plies
under pressure for 4 hours at a pressure of 400 psi. Figure 5.11 shows a 6 mm thick sample
formed by ITR.
Fig. 5.13. Unidirectional fabric, uncoated (a), coated with ATSP solution (b), cured lamina (c).
Prior to the VARTM process, the vacuum bagging techniques were tested to fabricate an
ATSP/C lamina using an ATSP prepreg. A 0.32 cm thick Aluminum tool plate (12" x 12") was
utilized as the base of the setup. Kapton polyimide bagging material and high temperature
sealant (A-800-3G, from Airtech) were used to seal the bag that contained the fiber preform,
PTFE coated fiberglass (from CS-Hyde) and breather material. A breather cloth (Airweave-
UHT-800) was also used to allow air and volatile escape (Figure 5.14).
(a) (b) (c)
144
Fig. 5.14. Vacuum bagging sample, after pulling vacuum.
The fiber-reinforced composite panel was cured at 202 °C for 30 min, 270 °C for 1 hour and
finally at 330 °C for two hours by applying 26 in Hg vacuum. The obtained cured lamina was
fully impregnated with ATSP and the resin content of cured composite was about 44%, but the
void content was about 23%. Vacuum bagging of ATSP impregnated weave fabric was also
explored, which resulted in void content of 12% (Figure 5.15). This is a reasonable value
considering that the fiber weave employed in this experiment was extremely loose (areal density
of 197 g/m2
[5.81 oz/sq. yd.]) and generally produced a high surface roughness of approximately
~100 μm of depth with lateral features of one millimeter in dimension. As the measurement of
this density relied on a micrometer with a 0.25” flat sampling surface, the thickness measurement
did not capture these depth features and the total sample thickness was only 0.9 mm, the surface
roughness was likely a significant component of total thickness and so the 12 vol% of void
content serves as an upper estimate.
145
Fig. 5.15. Weave fabric, uncoated (a), and cured lamina (b).
VARTM process was also conducted using the same unidirectional carbon fiber mat and
oligomer solutions used for prepreg route. C1A1 oligomer solution was heated to 70 °C to yield
viscosity of 50 cP. Figure 5.16 shows the values of C1A1 viscosity at different temperature
measured using a spindle type viscometer (Cole-Parmer, 98936 series). These values are stable
indefinitely as the cure reaction does not substantially advance at these temperatures. An
Aluminum tool plate was used as the base of the VARTM setup and Aluminum screen mesh as
the flow media. The high temperature materials that used for vacuum bagging process was
utilized for VARTM process (Figure 5.17). Tooling with fiber preforms (of [0/90]2 layup) inside
the polyimide bagging was then attached to vacuum apparatus and the heated container of
oligomer solution. The valve connecting the apparatus to the vacuum pump was then opened to
allow the resin solution to be drawn into the preform.
5
cm 5
cm
146
Fig. 5.16. Viscosity values of 50% C1A1 solution at different temperature, using LV-1 spindle
Fig. 5.17. VARTM setup sample, before addition of final Kapton sealant layer.
147
We were not successful in completely wetting out the fibers with ATSP resin during VARTM
process. This was determined by observing the weight increase after 1 and then 2 hours of
infiltration time. After one hour of infiltration time, sample evidenced a 21wt% increase in resin
content, with visually more resin concentrated on in-flow side of preform. This quantity did not
appreciably change after 2 hours of infiltration time. This was due, firstly, to unsuitability of
available fiber preforms. Sikawrap 103C was observed to have extremely large tow sizes stitched
with nylon thread. The unidirectional preform had a relatively low permeability within the tows
that likely prevent successful infiltration of more viscous resins. Secondly, ATSP resin solutions
have an increase in viscosity as temperature is decreased towards room temperature. Employed
setup did not utilize a temperature zone at the preform. This configuration allowed the ATSP
solution to cool to room temperature thus allowing an undesirable increase in viscosity and
preventing successful infiltration. Rectification of this issue could include: a modest decrease in
resin loading within solution, use of fabrics known to be amenable from literature to be VARTM
appropriate, and implementation of a temperature zone at the preform.
To determine the volume fraction of fibers, resin, and voids (following ASTM D3171),
representative samples of C1A1/C and C2A2/C, made via prepreg route, were first weighed and
the overall density of the samples was determined. The specimen were then digested using
concentrated sulfuric acid heated to about 80 °C. The samples were left in the acid for 6 hours to
remove all of the resin from the fibers. The solution was then filtered to recover the fibers and
rinsed with copious deionized water followed by isopropanol before being dried in a convection
oven at 60°C. The fibers were finally weighed and the volume fraction of fibers, resin, and voids
calculated. The results for C2A2/C composite sample are as follows:
Reinforcement content, volume percent:
148
Matrix Content, volume percent:
Void Volume, in percent:
Where, Mi (initial mass of the specimen) = 1.1932 g, Mf (final mass of the specimen after
digestion) = 0.7614 g, ρr (density of the reinforcement) = 1.8 g/cm3 [22], ρc (density of the
specimen) = 1.61 g/cm3, and
ρm (density of the matrix) = 1.35 g/cm
3.
The void content of ATSP/C composites reported in Table 5.2. C1A1 and C2A2/C composites
had relatively low void volume. Results observed in Table 5.2 are explainable given (1) the
higher acetic acid yield of the CBAB resin (7.5wt% vs 6.5wt% for C1A1 and 5wt% for C2A2),
(2) the broader Tg curve for C2A2 (indicating greater compliance during cure) as compared to
C1A1 and both C1A1 and C2A2 significantly lower cured Tg as compared to CBAB (see Figure
5.18), and (3) practical reality that all specimens were cured to a maximum of 330°C due to
equipment limitations. This means that CBAB evolved volatiles that due to sample rigidity
during cure and subsequent ITR steps that were not fully removed.
149
Table 5.2. Fiber, resin and void volume for ATSP composites
Dynamic mechanical analysis (DMA) was performed in a TA Instruments DMA Q800 to obtain
the storage modulus and Tg of ATSP via a 3 °C/min temperature ramp with a 1 Hz oscillation.
Neat ATSP specimens were cut to 25 mm long by 7 mm wide by 1 mm thick and were loaded in
a tensile clamp configuration. Dynamic thermal mechanical data as the storage modulus, loss
modulus and tan δ as a function of temperature are shown in Figure 5.18. The glass-transition
temperature (Tg) reported based on the tan δ peak was about 239.3 °C for C1A1 and 253.5 °C
for C2A2, and 307 °C for CBAB. As expected, the Tg of the material decreased with an increase
in molecular weight. As molecular weight increases, the resultant crosslink density decreases
which will result in a lowering of the Tg.
Cryogenic dynamic thermal mechanical data as the tan δ as a function of temperature for C1A1
and C2A2 neat polymers at different frequencies are shown in Figure 5.19. The tan δ peak at
around 0 and -50 °C was observed for C1A1 and C2A2, respectively.
150
0 50 100 150 200 250 300
1
10
100
1000
Mo
du
lus (
MP
a)
Temperature (°C)
Storage Modulus (MPa)
Loss Modulus (MPa)
tan
0.0
0.1
0.2
0.3
0.4
0.5
tan
0 50 100 150 200 250 300 350
1
10
100
1000
Storage Modulus (MPa)
Loss Modulus (MPa)
tan
Mo
du
lus (
MP
a)
Temperature (°C)
0.0
0.1
0.2
0.3
0.4
tan
0 50 100 150 200 250 300 350 400
1
10
100
1000
Mo
du
lus (
MP
a)
Temperature (°C)
Storage Modulus (MPa)
Loss Modulus (MPa)
tan
0.0
0.1
0.2
0.3
0.4
tan
Fig. 5.18. DMA results for neat ATSP samples, (a) C1A1, (b) C2A2, and (c) CBAB.
-150 -100 -50 0 50 100 150
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
Ta
n D
elta
Temperature (°C)
1.00 Hz
1.80 Hz
3.20 Hz
5.60 Hz
10.00 Hz
17.80 Hz
31.60 Hz
56.00 Hz
100.00 Hz
Cooling
Heating
-150 -100 -50 0 50 100 150
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0.13
0.14
0.15
Ta
n D
elta
Temperature (°C)
1.00 Hz
1.80 Hz
3.20 Hz
5.60 Hz
10.00 Hz
17.80 Hz
31.60 Hz
56.00 Hz
100.00 Hz
Cooling
Heating
Fig. 5.19. Cryogenic DMA results for neat ATSP samples, (a) C1A1, (b) C2A2.
(a) (b)
(c)
(a) (b)
151
Thermogravimetic analysis (conducted on a TA Instruments TGA 2950) was also used to study
the thermal stability of the C1A1/C, C2A2/C and CBAB/C composites. ATSP composites were
cut and tested in the TGA under nitrogen and air at heating rates of 10 °C/min. Figure 5.20
shows the trace of the mass loss/temperature derivative curve as well as the total mass loss. The
composites were stable even above 400 °C which reiterates the high temperature stability of this
resin system. The weight loss was about 60, 67 and 62% for C1A1/C, C2A2/C and CBAB/C
composites at 800 °C, respectively in air and about 27% for all composites in nitrogen.
Fig. 5.20. TGA curves of C1A1/C, C2A2/C and CBAB/C composites in air and nitrogen.
152
Isothermal heat stability of ATSP at 371 °C was also characterized for C1A1/C, C2A2/C and
CBAB/C composite. The ATSP/C samples were ramped up to 371 °C at a rate of 10 °C/min and
then held isothermally at that temperature for 1 hour in an air atmosphere. The weight loss at this
temperature was about 3.52, 3.05, and 1.33% for C1A1/C, C2A2/C and CBAB/C composites,
respectively (Figure 5.21). Note that in the temperature range of testing, carbon fiber is stable
and should not contribute to the weight loss. Some high temperature polyimides mass loss are
presented in Table 5.3. As shown, CBAB/C lost just 1.33% of its weight was. Even though
C1A1 and C2A2/C mass loss at 371 °C was about 1 % higher than that of polyimides (Table 5.3)
[7], one should note that the polyimides were cured at 371 °C for one hour during processing
while ATSP final cure temperature was at 330 °C.
Table 5.3. Cured Tg after 1 hour at 371°C and mass loss.
153
0 10 20 30 40 50 60 70 80 90 100
94
95
96
97
98
99
100
C/C1A1
C/C2A2
C/CBAB
We
igh
t (%
)
Time (min)
0
100
200
300
400
Te
mp
era
ture
(C
)
Fig. 5.21. Isothermal heat stability of ATSP/C composites.
CTEs for composite specimens was measured on a dilatometer. ATSP (C1A1/C and C2A2/C) and 8551-
7/C (provided by NASA) samples were sectioned into 3 x 3 x 25 mm blocks. The specimen was placed in
a dilatometer between a quartz fixture with a constant normal force that held onto the sample. The
temperature was increased at a rate of 5°C/min from RT to 300 °C. Change in displacement with
increasing temperature curves for ATSP and NASA samples are shown in Figure 5.22. CTE for both
longitudinal and transverse sections of ATSP composites as well as [0/90]2s layup of ATSP and 8551-7/C
are reported in Table 5.4.
154
0 50 100 150 200 250 300
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
% L
ine
ar
Ch
an
ge
(P
LC
)
Temperature (C)
8771/IM7 [0/90]7,alt
C1A1/103C [0]8
C1A1/103C [90]8
C1A1/103C [0/90]8,alt
C2A2/103C [0]8
C2A2/103C [90]8
C2A2/103C [0/90]8,alt
Fig. 5.22. Change in displacement with increasing temperature plot for ATSP and NASA composites.
Table 5.4. CTE of ATSP and 8551-7/C composites
The CTE of 8551-7/C is higher than that of ATSP/C composites. We expect that the residual
stresses induced in ATSP/C composites will be lower than that of 8551-7/C, thus ATSP/C can
help in providing reliable composite structures when used over a range of temperatures.
Four point flexural tests of the ATSP composites were conducted according to ASTM D7264 at
room temperature with the loading rate of 2 mm/min, as shown in Figure 5.23. This procedure
involved arranging cured lamina in unidirectional configurations [0°] which were then solid-state
bonded using ITR at a temperature of 330 °C for 4 h under a compressive pressure of 100 psi
under vacuum. Finished laminates were then cut into specimens of 130 mm in length for testing.
155
The width and thickness of samples are shown in Table 5.5. Flat rectangular specimens, 8-ply
unidirectional fabrics, of (C1A1/C and C2A2/C) composites were loaded in 4-point bending
Instron machine.
Fig. 5.23. Flexural testing being conducted.
Flexural strength, σ, and modulus, Ef, of the composites were calculated using the following
equations:
where, P is the breaking force of the specimen, L is the span support (105 mm), b is the width
and h is the thickness.
The flexural modulus, Ef, is calculated by drawing a tangent to the steepest initial straight-line
portion of the load deflection curve and using the following equation:
where m is the slope of tangent of the initial straight-line portion of the load deflection curve.
156
A minimum of five tests were conducted for each chemistry, and the average values and
variations are presented in Table 5.5. The stress–strain curve from sample # 1 is also shown in
Figure 5.24. Note that the load increases linearly until the failure point.
Table 5.5. Flexural properties of C1A1/C and C2A2/C composites.
Sample Resin Flexural Stress Flexural Modulus b h span/thickness
ID# Chemistry (MPa) (GPa) (mm) (mm) L/h
1 C1A1 993.7 120.5 13.03 4.00 26.25
2 C1A1 1161.0 114.1 12.10 3.97 26.45
3 C1A1 1262.0 127.7 12.88 3.88 27.06
4 C1A1 877.9 108.7 12.76 3.80 27.63
5 C1A1 1111.1 125.2 12.50 3.90 26.92
Mean C1A1 1081.1 119.2 12.7 3.9 26.9
Standard deviation C1A1 149.1 7.8 0.4 0.1 0.5
7 C2A2 879.0 119.4 12.22 2.25 46.67
8 C2A2 712.9 108.1 13.22 2.97 35.35
9 C2A2 660.9 109.8 13.80 2.82 37.23
10 C2A2 759.5 108.3 13.10 2.20 47.73
11 C2A2 770.9 115.4 12.25 2.23 47.09
12 C2A2 841.1 109.2 14.11 2.34 44.87
Mean C2A2 770.7 111.7 13.1 2.5 43.2
Standard deviation C2A2 80.3 4.6 0.8 0.3 5.4
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
0
100
200
300
400
500
600
700
800
900
1000
1100
Fle
xu
ral str
ess (
MP
a)
Flexural Strain (%)
Echord
f =
= 120.5 GPa
Maximum stress = 993.7 MPa
Maximum strain = 0.823 %
Fig. 5.24. Stress-strain curve from flexural test (sample #1).
157
Following the testing, all specimens were inspected. As shown in Figure 5.25, the most common
failure in nearly all composites was compressive local buckling at the outer (top) surface
between the loading noses. The C1A1/C composites evidenced a flexural strength and moduli of
1081.1±149.1 MPa and 119.2±7.8 GPa, respectively which is comparable or better than that of
polyimide and epoxy/carbon unidirectional composites available in literature [7, 23-24].
Therefore, we can gain the qualitative observation that out-of-plane shear moduli and strength
should be fairly high. This becomes an effective demonstration of the ability of fully-cured
ATSP laminae to be bonded together via interchain transesterification reactions (ITR) into an
effective multi-ply laminate.
C2A2 composites evidenced a somewhat lower flexural strength than C1A1-based composites
for one of two reasons: produced C2A2 composites were generally thinner in the thickness
dimension that C1A1 composites while the test was still conducted with a specified span length,
resulting in a change in span ratio (span length versus thickness, Table 5.5). This may have
created an earlier onset of failure events due to an increase in compressive stress concentration at
the upper surface (inside curvature) due to overall great deflection of the C2A2 samples during
testing.
158
Fig. 5.25. C1A1 (top) and C2A2 (bottom) specimens after flexural tests.
The fracture toughness and damage resistance of ATSP composites was characterized as shown
in Figure 5.26. Composite specimens, 20 mm wide, 3 mm thick and at least 125 mm long, with
[0/90/0]s layup pattern was used for testing ATSP matrices. A precrack was introduced in the
sample during the ITR cycle by placing a PTFE ply between the lamina. The samples were
loaded under tension and the two faces were pulled apart at a constant rate of 2 mm/min. The
fracture toughness of C1A1/C and C2A2/C was determined by loading a double cantilever beam
sample under tension at -90, RT and 100 °C. Samples were allowed to equilibrate with their
environment for 5 minutes prior to testing. A representative curve of the load vs displacement is
shown in Figure 5.27. For each sample the load was calculated as the average of the loads for up
to 90 mm of delamination length.
The GIc values of the various samples were calculated. The results are summarized in Table 5.6,
and a representative Mode I fracture toughness versus delamination length curve for C2A2/C
composite is shown in Figure 5.28.
159
Fig. 5.26. Fracture toughness testing being conducted.
0 10 20 30 40
0
10
20
30
40
50
Lo
ad
(N
)
Crosshead displacement (mm)
Applied load
0
10
20
30
40
50
60
70
80
90
100
Delamination length, a
De
lam
ina
tio
n le
ng
th,
a (
mm
)
Fig. 5.27. Representative load vs displacement and delamination length vs displacement curves
for C2A2/C composite obtained during fracture toughness testing a 100 °C.
160
Table 5.6: Interlaminar fracture toughness properties of ATSP/C composites
50 60 70 80 90 1000
100
200
300
400
500
600
700
800
Mo
de
1 inte
rlam
ina
r fr
actu
re to
ug
hn
es,
G1
C (
J/m
2)
Delamination length, a (mm)
Direction of crack advancement
Fig. 5.28. Representative Mode I fracture toughness vs delamination length curve for C2A2/C
composite at 100 °C.
Calculated GIc values were obtained via modified beam theory (MBT) as given in section 13.1.1
in ASTM D5528 with modifications for compliance and loading block geometry recommended
in Annex A1. Results indicate that both C1A1 and C2A2-based composites based on laminae
joined via solid-state ITR reactions evidenced fracture toughness values comparable or better
161
than an assortment of commercially-produced epoxy prepreg composites reviewed by Barikani et
al [116]. This demonstrates the viability of solid-state bonding via the ITR mechanism for
fabrication of composite structures. Both C1A1 and C2A2, as well, evidenced no significant
decline at cryogenic and elevated temperatures, indicating the potential for their use within
elevated temperature and cryogenic conditions. There was a modest enhancement of properties
observed for C2A2 (which was based on liquid crystalline oligomers and which continues to
evidence birefringence in the fully cured state) versus amorphous C1A1. This is thought to have
physical origin in one or both of (1) elevated void content for C1A1 composites, or (2) local
microfibrillation and enhanced cusping in C2A2 as suggested by the persistence of markers of
liquid crystallinity from the oligomeric to the fully cured state and ex situ observations of
fracture surfaces of C2A2 relative to C1A1.
-100 -50 0 50 100
0
100
200
300
400
500
600
700
800
900
1000
Mo
de
I inte
rlam
ina
r fr
actu
re to
ug
hn
ess G
1C
(J/m
2)
Temperature (C)
C1A1
C2A2
Fig. 5.29. Temperature dependence of Mode I interlaminar fracture toughness of ATSP/C
composites
162
Short beam bending tests were used to introduce interlaminar cracks of the ATSP/C [0/90]4s
composites. ASTM D2344 was closely followed during testing. Samples with dimensions of
about 36 mm x 12 mm x 6 mm were tested on an Instron machine with a 50 kN load cell at -90,
RT and 180 °C. A span length of 24 mm was used for all the testing and the experimental setup
is shown in Figure 5.30. ATSP [0/90]4s composites were loaded at a rate of 1 mm/min until
interlaminar cracks developed. Load versus displacement curves for a C1A1/C and a C2A2/C
sample at RT was plotted (Figure 5.31). The point of appearance of interlaminar cracks or
reduction of the load by 20% was noted as the interlaminar shear strength (ILSS) of the sample.
The summary of SBS experiments for 5 samples of each chemistry is shown in Table 5.7. These
values are comparable or better than polyimide and epoxy composites [13-15,116] of similar
layup.
Fig. 5.30. Short Beam Shear testing being conducted.
163
Fig. 5.31. Representative displacement vs load curve for C1A1/C and C2A2/C composites at
room temperature.
Table 5.7. Summary of ATSP/C SBS experiments at different RT.
Interlaminar shear strength of ATSP composites was also measured at different temperatures
(Table 5.8). It can be seen that there is an interesting phenomena evident in the difference
between the RT properties of the amorphous C1A1 resin and the liquid crystalline C2A2 resin. In
the first part, the amorphous ATSP resin has a strongly enhanced ILSS value relative to
164
published values for non-unidirectional (parallel to the span direction) SBS experiments. These
values were then observed to decrease significantly in cryogenic conditions. On the other hand,
C2A2 had values in line with its low degree of unidirectionality (<50%) and in fact had enhanced
ILSS values at cryogenic conditions. Its high temperature ILSS values were as well qualitatively
in line with relative decline in storage modulus observed in the DMA curve.
Table 5.8. Summary of ATSP/C [0/90]alt SBS experiments at different temperatures.
A common damage type of polymer matrix composites is matrix cracking, which may lead to
delamination and ultimately failure. Thermal residual stresses due to exposure time at different
temperatures are among the many factors that can influence the development of microcrack
damage. Even if microcrack density is not sufficient to diminish structural properties, the
presence of microcracks leads to the risk of permeation. Leakage of cryogenic fluids is among
the greatest concerns in the utilization of PMC’s for cryogenic-fuel tanks.
C2A2/C composite specimens were subjected to 170 h at 280 °C in air in a Lindberg tube
furnace. Samples tested in this manner were then subjected to short beam shear strength. A
165
representative displacement versus load curve for ATSP composites is plotted in Figure 5.32.
Interlaminar shear strength of C2A2/C composite was 16.05 ± 1.74 MPa after aging.
0.0 0.5 1.0 1.5
0
5
10
15
20
Str
ess (
MP
a)
Crosshead displacement
Point of interlaminar shear failure
Fig. 5.32. Representative displacement vs load curve for C2A2/C composite.
Isothermal aging conducted at 280 °C for 170 h indicates that C2A2/C composite display
complete property retention.
The thermal fatigue characteristics of ATSP/C composites, as compared to 8551-7/C composites
(provided by NASA), were studied by thermal cycling over a range of temperatures. Cured
ATSP/C laminas with [0/90]2s layup were cut into 4.5 x 4.5 cm squares to be compared to 8551-
7/IM7 carbon fiber composite.
The ATSP Thermal Cycling Machine (TCM) was built and installed (Figure 5.33). The TCM is
comprised of an upper square aluminum "oven" (heated by twin 240 Watt 3" x 12" strip heaters
and controlled by a single-input PID through twin solid state power relays), a sample holding
sled coupled to a drive motor system, a series of guide tubes, and a liquid nitrogen dewar. The
samples are held in a set of three waterjet-cut aluminum plates. The 304SS wire rope connecting
166
the reel to the sample sled assembly is 1/32" in diameter to minimize thermal mass and
maximize flexibility. The sled assembly is guided by four long 3/8" PTFE guide tubes, each
mounted on one end to a bottom plate and on the other end to the top plate shaft collars. The total
length of wire does not allow the sled to travel further down than a prescribed distance off the
bottom of the liquid nitrogen dewar.
Fig. 5.33. The ATSP Thermal Cycling Machine.
Four samples each of C1A1/C, C2A2/C and 8551-7/C ([0/90]2s) composites were cycled from
cryogenic bath to the oven. The sample were cooled in liquid nitrogen temperature (-196 °C) for
8 minutes and then heated up to the ambient oven temperature (35 °C) again for another 8
minutes.
To measure permeability of samples, each specimen was pressurized to 50 psi with nitrogen and
the pressure transducer in the leak tester (Figure 5.34) measured pressure changes. A drop in
pressure indicated a leak in the sample. Thermal cycling (-196⁰C to 35⁰C) of these of
167
ATSP/C specimens lead to the formation of microcracks, over time, formed a percolating
crack network from one side of the composite to the other, resulting in a gas permeable
specimen after 80 cycles. To quantify the healing ability of the ATSP/C composites, samples
were heated to 330 °C under 400 psi to induce ITR (as discussed earlier). Figure 5.35 shows the
permeability of samples after cycling as well as after heating under pressure. As seen, ATSP
shows a unique capability to repair microcracks upon application of heat and pressure.
Fig. 5.34. Pressure cell apparatus.
Fig. 5.35. Leak testing after 80 thermal cycles on ATSP composites, both as processed and
healed.
168
In another set of experiments, ATSP composites were fabricated using [0]7 layup and tight
weave E-glass fiber 7781 mat. Composites were then cycled from cryogenic bath to the oven.
The samples were cooled to liquid nitrogen temperature (-196 °C) in 5 minutes, then heated up to
the oven temperature at 90 °C, which accelerated heating process of the samples, thus samples
reached 35 °C in 5 minutes. Therefore, we managed to perform more cycling by decreasing the
half cycle time from 8 to 5 minutes.
After 300 cycles, none of the samples showed any leakage. Thermal fatigue testing suggests that
ATSP is resistant to microcracking. ATSP also shows a unique capability for healing of
interlaminar cracks on application of heat and pressure. These features, along with the high
strength-to-weight ratio, should make ATSP a prime candidate material to address a wide range
of current and future NASA missions and enabling technology for building lighter, highly
reusable, and more robust propellant tanks, thus reducing need for maintenance between flights.
The explanation for leaks seen in the first set of experiments for ATSP [0/90]2s composites can
be accounted for in the particulars of the weave of the 103C fiber mat used versus the 8551-7
prepreg-derived composite specimens used as comparison. The 103C mat received by ATSP in
2014 had a relatively large tow size compared to 8551-7 and this allowed (1) more substantial
net thermal stresses between tows which allowed more extensive inter-tow cracking and (2)
during cure, occasional gaps were opened during cure. This was not the case with the version of
unidirectional Sikawrap 103C received in 2009 and used previously and which possessed a far
smaller tow size. If one takes another example of even a very loose plain weave mat (see Figure
5.15), pressure cell testing evidenced a substantially reduced permeability despite a tow-tow
spacing of over a mm. Extending this to a very tightly woven glass fiber 7781/C1A1 and
169
7781/C2A2 composite we can also see that neither of these have leaked as well. This suggests
that the origins of the permeability in [0/90]2s ATSP composites used to date were most likely (1)
hand impregnation of the fiber mat prior to making composite and (2) large tow size of the fiber
mat. Future experiments along this line towards testing resin effect on composite permeability
should use fiber mats as similar as possible with only fiber sizing for compatibility with
particular resin acting as the only allowable variance.
170
CHAPTER 6
INTERCHAIN TRANSESTERIFICATION AS A SOLID-STATE
FABRICATION TOOL
6.1: Introduction and Background
Weight reduct ion is a benchmark that has always been crit ical to the aerospace and
automotive industries and metal replacement is often key to reducing weight, cost,
production t imes and processing cycles [110]. Polymers can be considered for metal
replacement for wide variety of applications in a range of uses such as
manufacturing equipment to automotive engines, aircraft components, oil and gas
process and extraction equipment, bushings, bearings, seals and gears. Here, we
outline results to date for a new technique to use interchain transesterificat ion for
solid-state fabricat ion of polymeric structures and propose several more. Fully
dense ATSP plates as well as bulk ATSP with different densit ies are fabricated by
mixing ATSP oligomers, curing to make foams, g rinding the cured material and
sintering the cured powders by applied heat/pressure. As well, rigid and
mechanically robust ATSP foams (from which the bulk plates are derived) are also
characterized in terms of thermal and mechanical properties for the fir st time.
Solid-state bonding of aromatic thermosetting copolyesters was previously
demonstrated by Frich [10,11] wherein two fully cured coated layers were bonded
to each other and demonstrated lap shear strengths compet it ive with commercial
state-of-art. It was also revealed by SIMS that the interdiffusion depth between the
171
cured layers (dist inguished by use of deuterium-tagged oligomers) was less than 50
nm. As this is significant ly below length scale understood to be needed for a purely
physical bond [18], it suggested a solid state chemical react ion at the bondlines.
Along with experimental demonstration of a chemical basis for sequence reordering
in hydroxybenzoic acid/hydroxynapthoic acid (HBA/HNA) copolymers [ 54],
interchain transesterificat ion react ions (ITR) were proposed as the operative
chemical mechanism for several related phenomena within the aromatic polyester
family. Following work utilized bonding schemes based on ITR to fabricate
cont inuous fiber composites [5] and polytetrafluoroethylene (PTFE)-filled
tribological materials [19,20,107] from ATSP synthesized as a cured powder via a
solut ion-suspension polymerizat ion. Here, we describe production of first ly a high
performance foam that is then recycled via grinding into a powder and sintered via
hot press into a fully dense (and presumably chemically cont iguous) bulk polymer
structure. Thermal and mechanical properties of the foam and bulk material,
including PTFE are characterized and reported.
Among the various solid state adhesive or reversible organic systems
(encompassing materials capable of schemes based on polymer shape memory [ 26-
28], transreact ions [4,96,112], hydrogen bonding [29], and others [24,30-31]), all-
aromatic ATSP appears to have the highest glass transit ion so far reported in
literature (310°C) for any material that has access to such a scheme.
172
Four addit ive manufacturing schemes and one welding scheme ut ilizing ITR are
also proposed for future work.
6.2: Synthesis and Properties of Foamed ATSP
Dynamic mechanical analysis (DMA) was performed in a TA Instruments DMA
Q800 to obtain the storage modulus and Tg of ATSP via a 3 °C/min temperature
ramp with a 1 Hz oscillat ion. Neat ATSP specimens were cut to 25 mm long by 7
mm wide by 1 mm thick and were loaded in a tensile clamp configurat ion.
Thermal stabilit y and isothermal heat stability of ATSP were characterized in a TA
Instruments TGA 2950 for fully dense C2A2 and CBAB structures. The ATSP
samples were heated to 600 °C with the heating rates of 10 °C/min for the former
and ramped up to 371 °C at a rate of 10 °C/min and then held isothermally at that
temperature for 3 hours in an air atmosphere, for the later experiments.
Coefficient of thermal expansion (CTE) of ATSP materials was measured on a
Orton Dilatometer. ATSP samples were sectioned into 3 x 3 x 25 mm blocks. The
specimen was placed in a between quartz fixtures with a constant normal force that
held onto the sample. The temperature was increased at a rate of 3°C/min from RT
to 300 °C. Observed percent linear change values below glass transit ion temperature
were used to develop coefficients of thermal expansion.
Compression tests on the ATSP foam and fully dense structures were conducted
according to ASTM D1621-10 and ASTM 695-10, respect ively at room temperature
with the loading rate of 1 mm/min in an Instron mechanical test ing machine. For
173
foam structures, test specimen were square in cross sect ion, 25.8 cm2 in area and
2.54 cm thick, and for the fully dense materials, cylindrical samples (5 mm in
diameter and 10 mm in length) were obtained.
The tensile strength of ATSP foams and fully dense materials was determined with
guidance from ASTM D638-14. The Type I sample specimen with the dumbbell
shape and thickness of 7 mm were used for fully dense material and tested in
tension in an Instron mechanical test ing machine. The straight gage sect ion has a
length of 50 mm and a width of 13 mm. For the foam structures, Type IV specimen
with gauge widths of 4 mm and thickness of 2 mm were used.
Samples for scanning electron microscopy were prepared via sect ioning fracture
surfaces from specimens and sputtering on a conduct ive layer of Au -Pd. A Hitachi
S-4700 scanning electron microscope (SEM) was used to obtain the micrographs.
ATSP foam structures were obtained by simply mixing two matched set oligomer
powders together (C2A2 and CBAB) via vigorous shaking in a container by hand
followed by applicat ion of heat. C2 and A2 oligomers were mixed C2:A2 at 1.1:1
weight ratio while CB and AB oligomers were mixed CB:AB at 1:1 and cured at 200
°C for 1 hr, 270 °C for 2 hr followed by 330 °C for 3 hr. Curing was performed
under vacuum to reduce oxidat ion. Mold release solid is necessary and only PTFE
was found to actually be compat ible and provide adequate release. Complex solids
of revolut ion were easily fabricated.
174
Acet ic acid is the by-product upon crosslinking of carboxylic end group and
acetoxy end group oligomers. The low density/ high strength closed -cell foamed
structures of ATSP were obtained by evolution of acet ic acid (Figure 6.1). Product
foam density was about 0.36 g/cm3
by water displacement of faced (uncut) foam,
which was generally water-tight. The foam exhibited high thermal stability up to
250 °C as shown in Figure 6.2.
Figure 6.2 shows thermal stability of ATSP foams ramped up and held for 8 hours
at 100, 150, and 250 °C and 4 hours at 200 °C in air. The samples evidenced no
weight loss at 100 and 150 °C, with a weight loss of 0.003% and 0.3% at 200 and
250 °C, respect ively.
Fig. 6.1. Photograph of foamed ATSP sample.
1 cm
175
0 100 200 300 400 50099.60
99.65
99.70
99.75
99.80
99.85
99.90
99.95
100.00
100.05
We
igh
t (%
)
Time (min)
100 C
150 C
200 C
250 C
Fig. 6.2. TGA isothermal hold experiment curves of ATSP foam in air.
The compression strength of ATSP foamed structures was determined following
ASTM D1621-10. The test specimen square in cross sect ion, 25.8 cm2 area and 2.54
cm thick, obtained and were tested under compression in an Instron mechanical
testing machine. A crosshead displacement of 2 mm/min was ut ilized and the
compressive strength vs strain is reported in Figure 6.3 The C2A2 foam structures
evidenced a yield strength of 3.62 ± 1.0 MPa, while the CBAB foam structures
showed a yield strength of 7.51 ± 1.0 MPa.
The tensile strength vs strain of C2A2 foam structure and CBAB foam structure
were plotted in Figure 6.4. Foam ult imate tensile strength, percent elongat ion, and
modulus were reported in Table 6.1. Tensile strength was reported as the breaking
load divided by sample cross-sect ional area, percent elongat ion will be determined
by extensometer, and modulus by dividing the difference in stress corresponding to
any segment of sect ion on this straight line by the corresponding difference in
strain in the init ial linear portion of the curve.
176
0 10 20 30 40 50 60
0
2
4
6
8
10
12
14
Co
mp
ressiv
e S
tre
ss (
MP
a)
Compressive Strain (%)
C2A2-Foam
CBAB-Foam
Fig. 6.3. Compressive stress vs strain of C2A2 and CBAB foam structures.
0 1 2 3 4
0
1
2
3
4
5
6
7
8
9
10
Te
nsile
Str
ess (
MP
A)
Tensile Strain (%)
C2A2-Foam #6
CBAB-Foam #2
Failure in gauge
Fig. 6.4. Tensile stress vs strain of C2A2 and CBAB foam structure s.
177
Table 6.1. Tensile and compressive propert ies of C2A2 and CBAB foam structures.
Material Density
(g/cm3) Compression
Strength at 10%
(MPa)
Compression
Strength at 20%
(MPa)
Ultimate
compressive
strength (MPa)
C2A2 Foam 0.36 1.5 3.0 8.0
CBAB Foam 0.38 2.0 4.2 7.0
Vespel SF
(Polyimide
Foam) [129]
0.3 ± 0.1 2.0 3.4 4.4
ATSP foams occupy a potent ially very interesting niche among polymeric foam
materials. CBAB foams demonstrated comparable mechanical properties in both
compression and tension as compared to Vespel SF [129] while not having the
limitat ion of 2 mm thickness imposed to Vespel foams due to processing demands.
Indeed, compression specimens ut ilized during this study were of significant ly
higher thickness than Vespel SF foams and even larger structures are seen as viable.
As well, commercially scaled insulat ing materials found in the aerospace industry
were found (Figure 6.25) to evidence quite oxidat ive stability in comparison to
ATSP foams. It is forseeable that ATSP foam material may be employed as low
density core materials in high performance applicat ions where a high glass
transit ion and enhanced mechanical properties in the core may be desirous.
Addit ionally, the relat ive modulus versus relat ive density (Figure 6.26) indicates
that ATSP should deform via idealized bending behavior.
178
6.3: Synthesis and Properties of Fully Dense ATSP Structures via ITR
ATSP foamed structure was ground and sieved to produce powders with controlled
particle size distribut ions in the range of <90 µm. A Col Int. Tech. FW 800 was
used for crushing and grinding the C2A2 foam structure and the produced particles
were automatically screened through mesh of 90 µm using a Retsch® Sieve Shaker
machine. Figure 6.5 shows part icle size dist ribut ion (weight and numbe r frequency)
of ground C2A2 and CBAB powder passed through 90 µm sieve.
Fig. 6.5. Particle size (weight and number) distribut ion of CBAB passed through 90
µm sieve as analyzed by micrographic image analysis.
179
For making fully dense ATSP, the fine cured powders (< 90 um) were loaded into a
(6 ¼” x 6 ¼”) mold and compressed at 1000 psi in a hot press under vacuum. The
sample was heated to 340 °C over 2 hr and then sintered at 340 °C for 2 hr (6hr for
CBAB). In another experiment the sample was sintered at 340 °C for 6 hr. The
final product was a fully consolidated part with the density of 1.32 g/cm3 and 1.27
g/cm3 for C2A2 and CBAB respect ively.
Dynamic mechanical analysis (DMA) was performed in a TA Instruments DMA
Q800 to obtain the storage modulus and glass-transit ion temperature (Tg) of ATSP
via a 3 °C/min temperature ramp with a 1 Hz oscillat ion. Neat ATSP specimens
were cut to 25 mm long by 7 mm wide by 1 mm thick and were loaded in a tensile
clamp configurat ion. Dynamic thermal mechanical da ta as the storage modulus, loss
modulus and tan δ as a funct ion of temperature are shown in Figure 6.6. The Tg
reported based on the tan δ peak was about 211 °C for C2A2, and 261 °C for CBAB.
As expected, the Tg of the material decreased with an increase in the number
average molecular weight between crosslinks M c and lower rigidity of backbone
units included within the oligomer set.
180
50 100 150 200 250 300 350
0.1
1
10
100
1000
Mo
du
li (M
Pa
)
Temperature (°C)
Storage Modulus (MPa)
Loss Modulus (MPa)
0.01
0.1
1
tan
tan
50 100 150 200 250 300 350
0.1
1
10
100
1000
Mo
du
li (M
Pa
)
Temperature (°C)
Storage Modulus (MPa)
Loss Modulus (MPa)
0.01
0.1
1
tan
tan
Fig. 6.6. DMA results for neat ATSP samples, (a) C2A2, and (b) CBAB.
Isothermal heat stability of ATSP at 371 °C was also characterized in a TA
Instruments TGA 2950 for fully dense C2A2 and CBAB structures. The ATSP
samples were ramped up to 371 °C at a rate of 10 °C/min and then held isothermally
at that temperature for 3 hours in an air atmosphere. The weight loss at this
temperature was about 8.17 % for CBAB and 6.66 %for C2A2, from (Figure 6.7).
0 100 200
90
92
94
96
98
100
Weig
ht (%
)
Time (min)
CBAB
C2A2
100
200
300
400
Tem
pera
ture
(C
)
Fig. 6.7. Isothermal heat stability of neat ATSP, C2A2 and CBAB.
(a) (b)
181
The compression strength of ATSP fully dense materials was determined with guidance from
ASTM D695-10. Cylindrical samples (5 mm in diameter and 10 mm in length) were machined
out of bulk ATSP specimens and tested under compression in an Instron mechanical testing
machine. The compressive strength is reported in Figure 10. The ultimate compressive strength
was 282.8 ± 31.4 MPa for C2A2 sintered for 2 hr, from Example 1.D, 333.7 ± 18.9 MPa for
C2A2 sintered for 6 hr, and 303.8 ± 11.1 MPa for CBAB sintered for 6 hr.
35 34 33 32 31 30 29 28
0
50
100
150
200
250
300
350
Co
mp
ressiv
e S
tre
ss (
MP
a)
POSN mm
C2A2-2h-1ksi
CBAB-6h-1ksi
C2A2-6h-1ksi
Fig. 6.8. The compression strength of C2A2 and CBAB fully dense structures.
The tensile strength of ATSP bulk materials was determined with guidance from ASTM D638-
14. The Type I sample specimen with the dumbbell shape and thickness of 7 mm were used for
fully dense material and tested in tension with an Instron mechanical testing machine. The
straight gage section had a length of 50 mm and a width of 13 mm.
182
The tensile strength vs strain results are reported in Figure 6.9 and 6.10 for fully dense C2A2 and
CBAB parts, respectively. Ultimate tensile strength, percent elongation, and modulus are
reported in Table 6.2. Tensile strength was reported as the breaking load divided by sample
cross-sectional area, percent elongation was determined by extensometer, and modulus by
dividing the difference in stress corresponding to any segment of section on this straight line by
the corresponding difference in strain in the initial linear portion of the curve.
0 1 2 3 4 5
0
10
20
30
40
50
60
70
80
90
100
C2A2-6hr
Te
nsile
Str
ess (
MP
a)
Tensile Strain (%)
Elastic Modulus = 3.0 GPa
Failure in gauge section
Fig. 6.9. The tensile strength vs strain of C2A2 fully dense structures.
183
0 1 2 3 4
0
10
20
30
40
50
60
70
80
90
100
Te
nsile
Str
ess (
MP
a)
Tensile Strain (%)
CBAB-6hr
Failure in gauge section
Fig. 6.10. The tensile strength vs strain of CBAB fully dense structures.
Table 6.2. Tensile properties of C2A2 and CBAB fully dense structures.
Formulation Tensile Stress (MPa) Modulus (GPa) Elongation (%)
C2A2 87.2 ± 1.5 3.2 ± 0.3 7.0 ± 2.0
CBAB 66.9 ± 1.6 4.3 3.0 ± 0.5
Material Density
(g/cm3) Compression Strength
(MPa) Ultimate Tensile strength
(MPa)
C2A2 1.32 283 87.2
CBAB 1.27 304 66.9
VICTREX® PEEK
450G
1.43 125 98
The coefficient of thermal expansion (CTE) for bulk ATSP specimens was measured
on an Orton Dilatometer. ATSP (C2A2 and CBAB) samples were sect ioned in to 3 x
3 x 25 mm blocks. The specimen was placed in a dilatometer between quartz
fixtures with a constant normal force holding onto the sample. The temperature was
184
increased at a rate of 2°C/min from RT to 200 °C. CTE for ATSP samples is
reported in Table 6.3.
Table 6.3. CTE of ATSP bulk materials.
Material CTE (1/K)
C2A2 38 x 10-6
CBAB 41 x 10-6
PEEK 450G 55 x 10-6
Vespel SP-1 54 x 10-6
Tribological test ing was conducted using a high-pressure tribometer (HPT) with a
pin-on-disk contact geometry. The HPT simulates typical operat ing condit ions
found in an air-condit ioning compressor. The disk samples were made of gray cast
iron with a hardness of 95 HRB, while the pins were cut from the rectangular
composite out of the mold. All experiments were performed at temperature of 25 °C
and 60 °C under a load of 155 N. In order to make polymer pins for the pin -on-disk
tribological tests, the ATSP fully dense specimen was machined down to pins with a
diameter of 6.35 mm and a height of 11 mm.
Table 6.4. Tribological propert ies of neat C2A2 and CBAB pins.
185
ATSP fabricated in this method was also produced with variable density. C2A2 and
CBAB foamed structures were ground and sieved to produce powders with
controlled part icle size distribut ions in the range of <90 µm and <250 µm. Powders
were produced by grinding ATSP foam using a Col Int. Tech. FW 800 grinder and
automated screening through mesh of 90 µm using a Retsch®
Sieve Shaker machine.
The larger powders were then sieved through 250 µm sieve.
Figure 6.11. Part icle size (weight and number) distribut ion of C2A2 and CBAB
passed through 250 µm sieves.
186
For the bulk C2A2 and CBAB materials, the cured powders were loaded into a (½”
x 2”) compression mold and put in a hot press under vacuum. Th e samples were
heated to 340 °C over 1.5 hr and then sintered at 340 °C for 0.5 hr with either no
applied pressure or 1000 psi. Table 6.5 shows density based on part icle size and
applied pressure. The compression strength of ATSP variable density materials was
determined with guidance from ASTM D695-10. Cylindrical samples (5 mm in
diameter and 10 mm in length) were machined and tested under compression in an
Instron mechanical test ing machine. The ult imate compressive strength vs density is
reported in Figure 6.12. This demonstrates a clear dependence between the
compressive strength and the density of the fabricated ATSP components.
Table 6.5. Density of bulk ATSP per powder particle size.
187
0.8 0.9 1.0 1.1 1.2 1.3 1.4
0
50
100
150
200
250
300
C2A2
CBAB
Co
mp
ressiv
e S
tre
ng
th (
MP
a)
Density (g/cm3)
Figure 6.12. The ult imate compression strength vs density.
For the bulk C2A2/PTFE and CBAB/PTFE materials, the cured C2A2 and CBAB
powders were mixed with PTFE powder, respect ively, via vigorous shaking in a
container by hand. ATSP/PTFE of 75/25 wt% and 95/5 wt% were loaded into a (2”
x 2”) compression mold and pu t in a hot press under vacuum and 1000 psi pressure.
The samples were heated to 340 °C over 2 hr and then sintered at 340 °C for 6 hr.
The final products were fully consolidated parts with density of 1.52 g/cm3
and 1.50
g/cm3
for C2A2/PTFE (75/25) and CBAB/PTFE (75/25) composites, respect ively,
and 1.37 g/cm3
for both C2A2/PTFE (95/5) and CBAB/PTFE (95/5) composites.
The compression strength of ATSP fully dense materials was determined with guidance from
ASTM D695-10. Cylindrical samples (5 mm in diameter and 10 mm in length) were machined
and tested under compression in an Instron mechanical testing machine. The compressive
strength is reported in Figure 15. The ultimate compressive strength was 84.04 ± 1.54 MPa and
126.38 ± 9.20 MPa for C2A2/PTFE (75/25) and CBAB/PTFE (75/25) fully dense composites.
188
41.0 41.5 42.0 42.5 43.0 43.5 44.0 44.5 45.0
0
20
40
60
80
100
120
140
Com
pre
ssiv
e S
tree (
MP
a)
POSN mm
C2A2:PTFE-75:25
CBAB:PTFE-75:25
Figure 6.13. The compression strength of CBAB/PTFE and C2A2/PTFE fully dense
structures.
Dynamic mechanical analysis (DMA) was performed in a TA Instruments DMA
Q800 to obtain the storage modulus and T g of ATSP/PTFE via a 3 °C/min
temperature ramp with a 1 Hz oscillat ion. Neat ATSP specimens were cut to 25 mm
long by 7 mm wide by 1 mm thick and were loaded in a tensile clamp configurat ion.
Dynamic thermal mechanical data as the storage modulus, loss modulus a nd tan δ as
a funct ion of temperature are shown in Figure 16. The glass -transit ion temperature
(Tg) reported based on the tan δ peak was about 218 °C for C2A2/PTFE (75/25), and
273 °C for CBAB/PTFE (75/25).
189
Storage modulus is an important parameter for t he rigidity of materials. Figure 16
shows a very similar storage modulus curve to the pure ATSP. As seen, the storage
modulus of the 75:25 composite is slight ly lower than the neat ATSP sample over
the ent ire temperature range
50 100 150 200 250 300 350
0.1
1
10
100
1000
Storage Modulus (MPa)
Loss Modulus (MPa)
Storage Modulus (MPa)
Loss Modulus (MPa)
Mo
du
li (M
Pa
)
Temperature (°C)
0.01
0.1
1
tan
tan
tan
Neat CBAB
CBAB/PTFE 75/25
50 100 150 200 250 300 350
0.1
1
10
100
1000
Mo
du
li (M
Pa
)
Temperature (°C)
Storage Modulus (MPa)
Loss Modulus (MPa)
Storage Modulus (MPa)
Loss Modulus (MPa)
Neat C2A2
C2A2/PTFE 75/25
0.01
0.1
1
tan
tan
tan
Figure 6.14. (Upper) DMA results for neat ATSP and ATSP/PTFE composite samples. (Lower)
Ramped temperature TGA results in air and nitrogen for neat C2A2 and CBAB.
Tribological test ing was conducted using a high-pressure tribometer (HPT) with a
pin-on-disk contact geometry. The HPT simulates typical operat ing condit ions
found in an air-condit ioning compressor. The disk samples were made of gray cast
190
iron with a hardness of 95 HRB, while the pins were cut from the rectangular
composite out of the mold. All experiments were perfor med at temperature of 25 °C
and 60 °C under a load of 230 N. The results are shown in Table 6.6.
In order to make polymer pins for the pin-on-disk tribological tests, the ATSP/PTFE
composite was machined down to pins with a diameter of 6.35 mm and a heig ht of
11 mm.
Table 6.6. Tribological properties of C2A2/PTFE and CBAB/PTFE pins.
Figure 6.15 shows the COF versus wear rate for neat C2A2, CBAB along with
C2A2/PTFE and CBAB/PTFE composites at 25 and 60 °C. Where A1, A2, A3
denote pure C2A2, C2A2/PTFE (95:05), and C2A2/PTFE (75:25), respect ively, and
B2 and B3 denote CBAB/PTFE (95:05) and CBAB/PTFE (75:25), respect ively.
191
Figure 6.15. COF versus wear rate for neat C2A2/ PBAB and C2A2/PTFE and
CBAB/PTFE composites pins.
Within the same temperature: higher percentage of PTFE results in lower COF and
lower wear rate. For the 25% PTFE composites, higher temperature results lower
COF but higher wear rate. The 5% PTFE composites, higher temperature results in
lower COF and lower wear rate. CBAB/PTFE composites, in same temperature and
concentration, have higher COF and lower wear rate compared with C2A2
composites.
192
Table 6.7. Network parameters and observed properties
Material Branch
density (xc) Average Flory branch
coefficient (αAVE) (Mc)n
(g/mol) Tan
δ (°C)
5% Weight loss
Temperature (°C)
C1A1
[19,20] 0.1436 0.5692 856.50 268 410
C1A1
[48]
0.1436 0.5692 856.50 224 450
C2A2 [48]
0.0718 0.4758 1692.13 209 --
C2A2 0.0718 0.4758 1692.13 211 405
CBAB 0.1333 0.45588 1112.33 261 460
The network parameters xc and (Mc)N are defined earlier in Chapter 2. The Flory branch
coefficient for each produced oligomer was calculated as [50,51]:
where r is the init ial formula rat io of carboxylic acid to acetoxy gr oups prior to
oligomerizat ion, ⍴ is the formula rat io of carboxylic acids belonging to the
branching units (trimesic acid) to the total number of carboxylic acids prior to
oligomerizat ion. Extent of react ion for acetoxy and carboxylic groups was
represented by pA and pC, respect ively. F lory coefficients for acetoxy-capped
oligomers (A1, A2, AB, AB2) were calculated using the equat ion for ɑ on the left
and carbolxylic acid-capped oligomers (C1, C2, CB, CB2) using the equat ion on the
right. These carries the assumption of complete react ion of the counterpart
funct ional groups during oligomerizat ion and that therefore pC and pA could be
assumed to be 1 after complet ion of the oligomerizat ion process described above.
193
Average Flory branch coefficient was computed as the number average of the
const ituent oligomers.
In this case, we can see that a higher crosslink density (as indicated by several
metrics in Table 6.7) in conjuct ion with observat ions of total elongat ion (Table 6.2)
and observat ions of fracture structures of neat ITR’d specimens seen in Figures 6.16
and 6.17, that this increases britt leness in exchange for enhancement of glass
transit ion temperature.
Well dispersed PTFE addit ives (as seen by fribrillar morphology across Figure 6.16
e & f, shows further that an ITR process is amenable to micron-scale addit ives and
produces structures with outstanding mechanical properties that may have ut ility in
ant i-wear applicat ions, as indicated in Table 6.6 and Figure 6.15.
194
Fig. 6.16. SEM images of (a, b) neat CBAB, (c, d) neat C2A2, (e) CBAB/PTFE (75:25), and (f)
C2A2/PTFE (75:25) after compression test.
195
Fig. 6.17. Fracture surfaces of neat bulk tensile specimens. (a) C2A2, (b) CBAB.
6.4: Proposed Concepts for Use of ITR as a Solid -state Fabrication Tool
Composite tank joints, especially bolted joints, are part icularly troubling area
predisposed to leaks. Based on the success of using ITR in as a solid -state joining
technique, we suggest the explorat ion of several techniques for solid state joining
that are amenable to implementat ion both in cont inuous operat ion and for joining
large components. Convent ional joining techniques for both thermoset and
thermoplast ic matrices rely on either adhesive or mechanical fastening. Both
techniques, though widely employed, have several limitat ions. These include
196
mismatches in CTE, high stress concentrations, and damage induced in drilling
operation for mechanical fastening. For adhesive bonding, producing bondlines of
appropriate strength and finding compat ible adhesives can be challenging in the
case of thermoplast ic resins and can require substant ial addit ional setting t imes in
either case. One solut ion to this is to employ composite welding or fusion bonding
as a joining operat ion. This has been substant ially explored in the case of
thermoplast ic matrices [16, 35-38] and aspects of this approach have been found to
be attractive to industry. For example, an earlier study by the Boeing Company
indicated that a labor savings of up to 61% could be afforded by implementat ion of
a welding approach as compared to a bolt ing approach [16].
Fig. 6.18. Fusion bonding classificat ion. Adapted from [16].
Techniques for welding of polymers can be divided into several categories based on
the method of heat induced at the bondline (Figure 6.18). One technique from each
category has been chosen to demonstrate applicability of this technique to aromatic
thermosetting copolyester resins. Earlier in the document and in prior research in
197
the Economy group [5,51], it was demonstrated that carbon fiber reinforced ATSP
single plies could be fabricated and then joined in the solid state by means of ITR
in a hot press environment. Results in Chapter 5 indicate that ATSP/C composites
produced with such a solid-state bonding method yielded interlaminar shear
strengths, flexural strengths and moduli, and mode I fracture toughness at the
bondline at least comparable to present ly employed resin systems. In the future, we
could demonstrate that techniques tradit ionally only employed on thermoplast ic
resins are now applicable to ATSP.
The High Power Direct Diode Laser (HPDDL), which is an ideal source for laser
heat treating, ultrasonic welder, which applies high frequency ultrasonic acoustic
vibrat ions, and/or localized induct ion heater will be examined to weld ATSP/C
composite parts. These processes eliminate the use of large heat ing ovens since only
certain areas of a part in large mold tools or engine parts need to be welded.
These localized heat ing processing making it possible to weld complicated
geometries by using ITR. Furthermore, laser processing and induct ion heat ing are
non-tact ile: therefore, no non-thermal forces are applied to the workpiece. As well,
prepreg would no longer have to be stored at structure integrat ion facility, great ly
simplifying equipment and vent ialat ion needs. These methods have high automation
potent ial for use in industrial manufacturing [33]. Thus if successful, they can be
used to simplify production processes for large composite sect ions of launch
198
vehicles as well as having impact in a broader aerospace manufacturing context if
successful.
Finally, the effect of welding on the strength of welded joints could be studied. Several
coupons will be cut according to Figure 6.19 then welded together. Based on the high observed
Mode 1 fracture toughness of ATSP composites produced by solid-state joining via the ITR
mechanism, tensile testing of welded ATSP composite specimens could be conducted following
ASTM D3039 [130]. Variations in notch length (L) could be explored to determine critical
joining length for weldable thermoset composite parts. When combined with present data on
fracture toughness of ATSP composites, this should present a compelling argument towards the
utility of continuous fiber thermoset composites joinable in the solid state.
Fig. 6.19. Proposed welding scheme for fully cured ATSP laminates.
199
We also suggest four addit ive manufacture schemes (Schemes 1 through 4). Scheme
1 would be an analogous process to that current ly employed for addit ive
manufacturing of titanium and other re fractory metals (select ive laser sintering -
SLS). With their common applicat ion in sintering of metals in mind, industrial
sintering lasers can easily impart a heat flux of up to 450°C [124] to a polymer and
so the exchange react ions at polymer surface and relaxat ion of the part icle itself can
happen extraordinarily fast.
Scheme 1:
1. Produce matched crosslinked COOH, AcO oligomer sets
2. Cure and/or partially cure
3. Grind/sieve to appropriate mesh
4. ↑Tg softens/ITR bonded layer repeat
Scheme 2 would be the most familiar approach for polymers although would require
a higher temperature nozzle system than convent ional polymers and would employ a
build zone preheated up to 270°C
Scheme 2:
1. Produce matched COOH, AcO linear oligomers sets
2. React/extrude into filament
3. Melts/cures cools layer repeat
200
Scheme 3 would be similar to Scheme 1 although would use observat ions from
Chapter 3 wherein it was observed that acet ic acid by-product gas would not be
trapped at small enough sect ion thicknesses. Subsequent to layer build, melted layer
would be cured.
Scheme 3:
1. Produce matched crosslinked COOH, AcO oligomer sets
2. Grind/sieve to appropriate mesh
3. Laser or electron beam
4. Melts cures layer repeat
Scheme 4 would ut ilize the ability of ATSP to undergo ITR between coated plates
[11,12] and cured laminae [5,6,8,51] under heat and pressure. This process would
allow use of arbitrary lamina as stock sheets so long as surfaces are capable of ITR
and a laser can cut through the sheet.
201
Fig. 6.20. Addit ive manufacturing scheme 4
This process would allow very rapid fabrication of laminates of complex geometry
and arbitrary laminate order. It may have impact on rapid fabricat ion of inject ion
molding cavit ies, microfluidics, hypersonic vehicles, general parts fabricat ion in the
aerospace and automotive industries.
Examples of plausible laminae include:
1. Woven carbon fiber/ATSP composite
2. Woven E-glass/ATSP composite
3. ATSP Foam layers
202
4. Filled or unfilled bulk ATSP layers
5. ATSP-coated ceramic layers
6. Both-side-coated metal sheets (ferrous, Al, Ti)
6.5: Conclusions
It has been demonstrated that ATSP foams and ITR’d material derived from the
foams possess attractive mechanical properties compet it ive with the high est values
available from the literature within their overall form-factor. (Figures 6.21-6.19). It
As well, Figure 6.24 demonstrates its amenability to precision machining. With
these various aspects in mind, it seems clear that ITR is an extraordinarily p owerful
tool for polymer structural fabricat ion.
Fig. 6.21. Ashby plot of compressive strength versus density. Adapted from [ 127].
203
Fig. 6.22. Ashby plot of tensile strength versus density. Adapted from [ 126]
204
Fig. 6.23. Ashby plot of Young’s modulus versus density. Adapted from [126].
205
Fig. 6.24. Bulk C2A2 material with basic machining implemented including
threading.
Fig. 6.25. ATSP foams in comparison to other state of art insulat ing and lightweight
materials
206
Fig. 6.26. Ashby schematic plot of relat ive density versus relat ive modulus with
posit ion ATSP foams indicated in red. Adapted from [128].
207
REFERENCES
1. A. Donald, Liquid Crystalline Polymers, Cambridge University Press, (2006).
2. J. Economy and K. Goranov, “Thermotropic Liquid Crystalline Polymers for High
Performance Applications” Adv. in Poly. Sci., 117, 222-255 (1994).
3. J. Economy, B. Nowak, “Polyester based on hydroxybenzoic acids”, US Patent
US3759870 A, (1970).
4. D. Frich, K. Goranov, L. Schneggenburger, and J. Economy, “Novel High-Temperature
Aromatic Copolyester Thermosets: Synthesis, Characterization, and Physical Properties”
Macromolecules, 29, 7734-7739 (1996).
5. A. Lopez and J. Economy, “Solid State Bonding of Graphite/Thermoset Composites via
lnterchain Transesterification Reaction (ITR)” Polym. Compos., 22, 444-449 (2001).
6. J. Economy, A. Polycarpou, and J. Meyer, “Polymer coating system for improved
tribological performance” US Patent 20130337183 A1, (2013).
7. F. Shi and J. Economy, “Moisture Transport Studies on Newly Developed Aromatic and
Aromatic / Aliphatic Copolyester Thin Films” J. Poly. Sci. B: Poly. Phys., 36, 1025–1035
(1998).
8. B. Vaezian, J. Meyer, C. Mangun, J. Economy, “Aromatic Thermosetting coPolyester
Composites for High Temperature and Cryogenic Applications” NASA SBIR Phase I
NNX14CM10P – Final Report (2014).
9. Z. Parkar, C. Mangun, D. King, T. Field, G. Sutton, J. Economy, “Ablation
characteristics of an aromatic thermosetting copolyester/carbon fiber composite” J.
Compos. Mater., 46, 1819-1830, (2012).
10. D. Frich, A. Hall, J. Economy, “Nature of adhesive bonding via interchain
transesterification reactions (ITR)”, Macromol. Chem. Phys. 199, 913-921, (1998).
11. D. Frich, J. Economy, K. Goranov, “Aromatic Copolyester Thermosets: High
Temperature Adhesive Properties”, Poly. Eng. and Sci, 37, 541-548, (1997).
12. K. Chuang, D. Revilock, J. Pereira, J. Criss, E. Mintz, “High Temperature Polyimide
Composites Fabricated by RTM: Characterization and Impact Testing”, SAMPE Journal,
49, 5 48-57, (2013).
13. S. Ghose, K.A. Watson, R.J. Cano, S.M. Britton, B.J. Jensen, J.W. Connell, H.M.
Herring, and Q.J. Lineberry., “High Temperature VARTM using LARC-PETI-9
polyimide resin”, High Perf. Polym., 21, 653-672, (2009).
14. R.J. Cano, B.W. Grimsley, B.J. Jensen, and C.B. Kellen. “Processing and
Characterization of PETI Composites Fabricated by High Temperature VARTM”, 36th
Int. SAMPE Conf., 459-68 (2004).
15. K. Chuang, J. Criss, E. Mintz, “Polyimide Composites Properties of RTM370 Fabricated
by Vacuum Assisted Resins Transfer Molding (VARTM)” NASA Technical Report
20110008609, (2011).
16. C. de Ruijter, E. Mendes, H. Boerstoel, S, Picken, “Orientational order and mechanical
properties of poly(amide-block-aramid) alternating block copolymer films and fibers”
Polym., 47, 8517-8526, (2006).
17. S. Bhama and S. Stupp, “Liquid Crystal Polymer-Carbon Fiber Composites: Molecular
Orientation” Poly. Eng. and Sci., 30, 228-234, (1990).
208
18. T. Chung, Z. Gurion, J. Stamatoff, “Induced Orientational Behavior of Liquid Crystal
Polymer by Carbon Fibers”, Polym. Compos., 6, 181-184, (1985).
19. N. Demas, J. Zhang, A. Polycarpou, J. Economy, “Tribological Characterization of
Aromatic ThermosettingCopolyester–PTFE Blends in Air Conditioning Compressor
Environment”, Tribol. Lett., 29, 253–258, (2008).
20. J. Zhang, N. Demas, A. Polycarpou, J. Economy, “A new family of low wear, low
21. coefficient of friction polymer blend based on polytetrafluoroethylene and an
aromaticthermosetting polyester”, Polym. Adv. Technol., 19, 1105–1112, (2008).
22. Z. Parkar, J. Meyer, J. Economy “Repair of Interlaminar Cracks using Interchain
Transesterification Reactions in an Aromatic Thermosetting Copolyester/Carbon Fiber
Composite”, J. Compos. Mater. submitted.
23. G. Gardiner. “Building TRUST in bonded primary structures”, Composites World,
http://www.compositesworld.com/articles/building-trust-in-bonded-primary-structures,
(2015)
24. X. Luo, K. Lauber, P. Mather, “A thermally responsive, rigid, and reversible adhesive,”
Polym., 51, 1169–1175, (2010).
25. V. Pisacane, “Fundamentals of Space Systems”, Johns Hopkins University Applied
Physics Laboratory Series in Science & Engineering, 424, (2005).
26. T. Xie, J. Hulway, X. Xiao, “Shape memory polymer and adhesive combination and
methods of making and using the same”, US 8685528 B2, (2009).
27. T. Xie, “Recent advances in polymer shape memory”, Polym., 52, 4985-5000, (2011).
28. H. Mengand G. Li, A review of stimuli-responsive shape memory polymer composites,
Polymer, 54, 2199-2211, (2013).
29. M. Bartlett, A. Crosby, High Capacity, Easy Release Adhesives From Renewable
Materials, Adv. Mater. 26, 3405-3409, (2014).
30. D. King, M. Bartlett, C. Gilman, D. Irschick, A. Crosby, “Creating Gecko-Like
Adhesives for ‘Real World’ Surfaces”, Adv. Mater., 26, 4345–4351, (2014).
31. M. Bartlett, A. Croll, D. King, B. Paret, D. Irschick, A. Crosby, “Looking Beyond
Fibrillar Features to Scale Gecko-Like Adhesion” Adv. Mater., 24, 1078-1083, (2012).
32. A. Yousefpour, M. Hojjati, J-P Immarigeon, “Fusion Bonding/Welding of Thermoplastic
Composites”, J. of Thermoplastic Compos. Mater., 17, 303-341, (2004)
33. I. Villegas, “High-speed spot welding of continuous fibre reinforced thermoplastic
composites”, CfK Valley Convention, (2012).
34. I. Villegas, H. Bersee, “Ultrasonic welding of advanced thermoplastic composites: an
investigation on energy-directing surfaces”, Adv. in Poly. Tech., 2, 112-121, (2010)
35. R. Rudolf, P. Mitschang, M. Neitzel, “Induction heating of continuous carbon-fibre-
reinforced thermoplastics”, Compos. A: App. Sci. and Manuf., 31, 1191-1202, (2000).
36. J. Economy, US Patent No. 5,439,541, (1995).
37. D. Frich and J. Economy, “Thermally Stable Liquid Crystalline Thermosets Based on
Aromatic Copolyesters: Preparation and Properties,” J. Polym. Sci. A: Polym. Chem., 35,
1061 (1997).
38. J. Hummel and P. Flory. “Structural Geometry and Torsional Potentials in p-Phenylene
Polyamides and Polyesters”, Macromolecules. 13 479-484, (1980).
39. K. Berger and M Ballauff. Structure and phase behavior of thermotropic poly(2-n-alkyl-
1,4-phenylene terephthalate)s, Molecular Crystals and Liquid Crystals, 157, 109-23,
(1988).
209
40. T. Chung, M. Cheng, G. Min, H. Suat, M. Jaffe; GW Calundann. Revisiting the
crystallization mechanism of vectra, a liquid crystal polymer, J. of App. Polym. Sci., 72-9,
1139-1150, (1999).
41. S. Bhama and S. Stupp. Liquid crystal polymer-carbon fiber composites. molecular
orientation, Polym. Eng. and Sci., 30-4, 228-234, (1990).
42. P. Mangalgiri, “Polymer-matrix Composites for High-temperature Applications”,
Defense Sci. J. 55, 175-187, (2005).
43. S. Bender, M.S. Thesis, University of Illinois, (2009).
44. A. Klepl, “Uber die Produkte der trocknen Destillation von Paraoxybenzoësäure”, J.
Prakt. Chem., 28, 193-219, (1883).
45. E. Fischer, “Über die Carbomethoxyderivate der Phenol-carbonsäuren und ihre
Verwendung für Synthesen. II”, Berichte der deutschen chemischen Gesellschaft, 42,
215-228, (1909).
46. R.Gilkey, J. Caldwell, “Polyesters of hydroxybenzoic acids”, J. Appl. Polym. Sci., 2,
198–202, (1959).
47. J. Economy and Z. Parkar, “High Temperature Aromatic Polyesters of p-Hydroxybenzoid
Acid and Their Copolyesters”, In “100+ Years of Plastics, Leo Baekeland and Beyond”,
7, 93-103, (2011).
48. D. Frich, “The Effect of Interchain Transesterification on the Development of Aromatic
Copolyesters”, University of Illinois at Urbana-Champaign, (1996).
49. J. Gooch, “Encyclopedic Dictionary of Polymers”, Springer, 182, (2010).
50. H. Duran, B. Yameen, M. Geuss, M. Kappl, M. Steinharte and W. Knollf, “Enhanced
interfacial rigidity of 1D thermoset nanostructures by interface-induced liquid
crystallinity”, J. Mater. Chem. C, 1, 7758-7765, (2013).
51. Z. Parkar, “Design of Unique Composites Based on Aromatic Thermosetting
Copolyesters”, University of Illinois, (2011).
52. Y. Kim, J. Economy, “The Degradation Process Observed during Step Annealing of
73/27 HBA/HNA Copolyester”, Macromolecules, 32, 2855-2860, (1999).
53. J. Economy, L. A. Schneggenburger, “Solid state interchain transesterification reactions
in macromolecules”, Macromolecular Symposia, 128, 1-12, (1998).
54. L. A. Schneggenburger, P. Osenar, and J. Economy, “Direct evidence for sequence
ordering of random semicrystalline copolyesters during high-temperature annealing”
Macromolecules, 30, 3754-3758, (1997).
55. T. Chung, Z. Gurion, and J. Stamatoff, “Induced orientational behavior of liquid crystal
polymer by carbon fibers”, Polym. Compos., 6, 181-184, (1985).
56. T. Davies, “Chemical Reactions of Polymers”, Interscience, 553-561, (1964).
57. P. Flory, “Principles of Polymer Chemistry”, Cornell University Press, (1953).
58. E. Bostick, “Chemical Reactions of Polymers”, Interscience, 515-528, (1964).
59. M. Berenbaum, “Chemical Reactions of Polymers”, Interscience, 528-548, (1964).
60. M. Stern and A. Tobolosky, Stress-Time-Temperature Relations in Polysulfide Rubbers,
J. Chem. Phys., 14, 93, (1946).
61. M. Stevens, “Polymer Chemistry: An Introduction”, 3rd Edition, Oxford University
Press, (1999).
62. J. McMurry, “Organic Chemistry”, 8th Ed., Brooks Cole, (2011).
63. P. Flory, “Kinetics of the Degradation of Polyesters by Alcohols”, J. Am. Chem. Soc., 62,
2255-2261, (1940).
210
64. W. Jackson and H. Kuhfuss, “Liquid crystal polymers. I. Preparation and properties of p-
hydroxybenzoic acid copolyesters”, J. Polym. Sci., Polym. Chem. Ed. 14, 2043-2058,
(1976).
65. J. Jin, S. Lee, H. Park and I. Kim, “Liquid Crystalline Properties of the Copolyesters
Prepared from Resorcinol, p-Hydroxybenzoic Acid, and Terephthalic Acid”, Polym.
J.,21, 615-622, (1989).
66. H. Ramjit, “The Influence of Stereochemical Structure on the Kinetics and Mechanism of
Ester-Ester Exchange Reactions by Mass Spectrometry”, J. Macomol. Sci.-Chem., A20,
659-673, (1983).
67. E. Tijsma, L. van der Does, A. Bantjes, and I. Vulic, “Interchange reactions with
polypivalolactone”, Makromol. Chem., 194, 305-319, (1993).
68. P. Flory, “Fundamental Principles of Condensation Polymerization”, Chem. Rev., 39,
137-197, (1946).
69. K. Yoda, “Catalysis of trans-esterification reactions”, Makromol. Chem., 136, 311-313,
(1970).
70. F. Hamb, “Copolyesters of glycols and bisphenols: A new preparative process”, J. Polym.
Sci., Polym. Chem. Ed., 10, 3217, (1972).
71. E. Tijsma, L. van der Does, A. Bantjes, I. Vulic, G. Buning, “Interchange reactions with
polypivalolactone”, Makromol. Chem., 194, 2807-2825, (1993).
72. P. Flory, “A Comparison of Esterification and Ester Interchange Kinetics”, J. Am. Chem.
Soc., 62, 2261-2264, (1940).
73. J. Kugler, J. Gilmer, D. Wiswe, H. Zachmann, K. Hahn, E. Fischer, “Study of
transesterification in poly(ethylene terephthalate) by small-angle neutron scattering”,
Macromolecules, 1116-1119, (1987).
74. V. Arrighi, J. Higgins, R. Weiss, A. Cimecioglu, “A small-angle neutron scattering study
of a semiflexible main-chain liquid crystalline copolyester”, Macromolecules, 25, 5297-
5305, (1992).
75. A. Muhlebach, J. Economy, R. Johnson, T. Karis, J. Lyerla, “Direct evidence for
transesterification and randomization in a mixture of homopolyesters of poly(4-
hydroxybenzoic acid) and poly(6-hydroxy-2-naphthoic acid) above 450C”
Macromolecules, 23, 1803, (1990).
76. E. George and R. Porter, “Novel approach for the processing of thermotropic liquid-
crystal polymers. 1. Thermodynamics and solid-state transesterification”
Macromolecules, 19, 97-105, (1986).
77. M. Stewart, A. Cox, D. Naylor, “Reactive processing of poly(ethylene 2,6-naphthalene
dicarboxylate)/poly(ethylene terephthalate) blends” Polym., 34, 4060-4067, (1993).
78. J. Economy, T. Gogeva, V. Habbu, “Liquid Crystalline Copolyesters as High
Temperature Adhesives for Aluminum” J. Adhesion, 37, 215-224, (1992).
79. J. Economy, and A. Andreopoulos, “Factors Which Influence the High Temperature
Adhesive Characteristics of Liquid Crystalline Copolyesters”, 40, 115-125, (1993).
80. C. Klapperich, K. Komvopoulos, L. Pruitt, “Nanomechanical properties of polymers
determined from nanoindentation experiments”, J. Tribol. 123, 624-631, (2001).
81. B. Briscoe, L. Fiori, E. Pelillo, “Nano-indentation of polymeric surfaces”, J. Phys. D:
Appl. Phys., 31, 2395-2405 (1998).
211
82. A. Strojny, X. Xia, A. Tsou, W. Gerberich, “Techniques and considerat ions
for nanoindentat ion measurements of polymer thin film const itutive
properties”, J. Adhes. Sci. Technol., 12, 1299-1321, (1998).
83. X. Xia, A. Strojny, L.E. Scriven, W.W. Gerberich, A. Tsou, C.C. Anderson,
Const itutive property evaluat ion of polymeric coatings using nanomechanical
methods, Mater. Res. Soc. Symp. Proc. 522 199-204, (1998).
84. N. Tayebi, A. Polycarpou, T.F. Conry, “Effect of substrate on determination of hardness
of thin films by nanoscratch and nanoindentation techniques”, J. Mater. Res. 19 1791-
1802, (2004).
85. J. Hay, “Measuring substrate-independent modulus of dielectric films by instrumented
indentation”, J. Mater. Res. 24, 667-677, (2009).
86. E. Escobar Nunez, S. Yeo, K. Polychronopoulou, A. Polycarpou, “Tribological study of
high bearing blended polymer-based coatings for air-conditioning and refrigeration
compressors”, Surf. Coat. Tech. 205, 2994-3005, (2011).
87. T. Solzak, A. Polycarpou, “Tribology of hard protective coatings under realistic operating
conditions for use in oilless piston-type and swash-plate compressors”, Tribol. T. 53, 319-
328, (2010).
88. W. Oliver and G. Pharr, “An improved technique for determining hardness and elastic
modulus using load and displacement sensing indentation experiments”, J. Mater. Res. 7,
1564-1583, (1992).
89. T. Ohmura, S. Matsuoka, K. Tanaka, T. Yoshida, “Nanoindentation load-displacement
behavior of pure face centered cubic metal thin films on a hard substrate”, Thin Solid
Films 385, 198-204, (2001).
90. M. Mata, O. Casals, J. Alcala, “The plastic zone size in indentation experiments: The
analogy with the expansion of a spherical cavity”, Int. J. of Solids and Structures, 43,
5994-6013, (2006).
91. A. Bolshakov, G. Pharr, “Influences of pileup on the measurement of mechanical
properties by load and depth sensing indentation techniques”, J. Mater. Res. 13, 1049-
1058, (1998).
92. S. Murov, “Properties of Organic Solvents”, http://murov.info/orgsolvents.htm, (2014).
93. D. Crawmer, Ed: J. Davis, “Handbook of Thermal Spray Technology”, ASM
International, 47-53, (2004).
94. D. Crawmer, Ed: J. Davis, “Handbook of Thermal Spray Technology”, ASM
International, 53-76, (2004).
95. I. Daniel, and O. Ishai, “Engineering Mechanics of Composite Materials” 2nd
Ed., Oxford
University Press, (2006).
96. F. Calleja, S. Fakirov, H. Zachmannm, Ed: S. Fakirov, “Transreactions in Condensation
Polymers”, Wiley-VCH, 429-473, (2008).
97. C. Mangun, Personal Communication, (2014).
98. F. Shi and J. Economy, “Aliphatic/aromatic copolyester thermoset adhesives: Synthesis
and characterization”, Polym. Eng. and Sci., 37, 549-558, (1997).
99. M. Guo, “Design of aromatic thermosetting polyester composites for thermally stable
devices”, University of Illinois, (2014).
100. F. Shi, J. Economy, “Thermomechanical studies on newly developed copolyester thin
films. Part I: Thermal expansion coefficients”, Polym. Eng. and Sci., 38, 499-504, (1998).
212
101. F. Shi, J. Economy, “Thermomechanical studies on newly developed copolyester thin
films. Part II: Thermal and intrinsic stress”, Polym. Eng. and Sci., 38, 971-975, (1998).
102. J. Selby, M. Shannon, K. Xu, J. Economy, “Sub-micrometer solid-state adhesive bonding
with aromatic thermosetting copolyesters for the assembly of polyimide membranes in
silicon-based devices, J. Micromech. And Microeng., 11, 672-685, (2001).
103. K. Xu, J. Selby, M. Shannon, J. Economy, “Adhesion mechanisms in the solid-state
bonding technique using submicrometer aromatic thermosetting copolyester adhesive”, J.
App. Polym. Sci., 92, 3843-3856, (2004).
104. Y. Huang, J. McCormick, J. Economy, “Adhesion of an aromatic thermosetting
copolyester with copper foils”, Polym. for Adv. Tech., 16, 1-5, (2005).
105. J. Zhang, A. Polycarpou, J. Economy, “An Improved Tribological Polymer Coating
System for Metal Surfaces”, Tribol. Lett., 38, 355-365, (2010).
106. K. Xu, J. Economy, “Hyperbranched Thermosetting Poly(imide-esters): Synthesis and
Properties”, Macromolecules, 37, 4146-4155, 2004.
107. Y. Huang, J. Economy, “Wear properties of UHMWPE/aromatic thermosetting
copolyester blends in unlubricated sliding”, Wear, 262, 943-948, (2007).
108. F. Shi, L. Schneggenburder, J. Economy, “New photoimageable dielectric insulating
copolyester thin films: Synthesis and characterization”, J. App. Polym. Sci., 63, 1199-
1211, (1997).
109. S. Ebnesajjad and R. Morgan, “Fluoropolymer additives”, Elsevier, (2012).
110. L. Carlsson, D. Adams, R. Pipes, “Experimental Characterization of Advanced
Composite Materials, 4th Ed.”, CRC Press, (2014).
111. L. McKeen, “Fluorinated Coatings and Finishes Handbook: The Definitive User's
Guide”, Elsevier, (2006).
112. M. Capelot , D. Montarnal , F. Tournilhac , and L. Leibler, “Metal-Catalyzed
Transesterification for Healing and Assembling of Thermosets”, J. Am. Chem. Soc., 134,
7664–7667, (2012).
113. W. MacDonald, A. McLenaghan, G. McLean, R. Richards, S. King, “A neutron
scattering investigation of the transesterification of a main-chain aromatic polyester”,
Macromolecules, 24, 6164-6167, (1991).
114. A. Padias and H. Hall, “Mechanism Studies of LCP Synthesis”, Polymers, 3, 833-845,
(2011).
115. X. Han, A. Padias, H. Hall, H. Sung, “Synthesis of thermotropic LCPs usingp-
methoxycarbonyloxy aromatic acids”, J. of Polym. Sci. Part A: Polym. Chem., 37, 1703-
1707, (1999).
116. M. Barikani, H. Saidpour, M. Sezen, “Mode-I Interlaminar Fracture Toughness in
Unidirectional Carbon-fibre/Epoxy Composites”, Iranian Polym. J., 11, 413-424, (2002).
117. K. Holmberg, A. Matthews: “Coatings Tribology: Properties, Mechanisms, Techniques
and Applications in Surface Engineering”, Elsevier Science, (2009).
118. S. Yeo, A. Polycarpou, “Fretting experiments of advanced polymeric coatings and the
effect of transfer films on their tribological behavior”, Tribol. Int. 79, 16-25 (2014)
119. A. D’Amore, A. Pompo, L. Nicolais, “Viscoelastic effects in poly(ether ether ketone)
(PEEK) and PEEK-based composites” Compos. Sci. and Tech., 41, 303-325 (1991).
120. S. Shankar, D. Karthikeyan, L. Dokos, K. Kannan, “Global Powder Coatings Market:
Demand for High Performance and Sustainable Products Steers Growth”, Frost &
Sullivan, (2014).
213
121. K. Almara, S. Saber-Samandari, C. Berndt, “Splat taxonomy of polymeric thermal spray
coating”, Surf. & Coat. Tech., 205, 5028-5034, (2011).
122. F. Sciance, “The Transition from HFC-134a to a Low-GWP Refrigerant in Mobile Air
Conditioners”, General Motors Public Policy Center, (2013).
123. B. Minor, D. Spatz, “HFO-1234yf Low GWP Refrigerant Update”, International
Refrigeration and Air Conditioning Conference, (2008).
124. I. Gibson, B. Stucker, D. Rosen, “Additive Manufacturing Technologies: 3D Printing,
Rapid Prototyping, and Direct Digital Manufacturing, 2nd
Ed.”, Springer, (2015).
125. D. Thomas, S. Gilbert, “Costs and Cost Effectiveness of Additive Manufacturing”, NIST
Special Publication 1176, (2014).
126. A. Lovatt, D. Shercliff, “Material selection and processing”, http://www-
materials.eng.cam.ac.uk/mpsite/, (2002).
127. J. Bauer, S. Hengsbach, I. Tesari, R. Schwaiger, O. Kraft, “High-strength cellular ceramic
composites with 3D microarchitecture”, Proceedings of the National Academy of
Sciences, 111, 2453-2458, (2014).
128. M. Ashby, “The properties of foams and lattices”, Phil. Trans. R. Soc. A, 364, 15-30,
(2006).
129. “Dupont Vespel SF Polyimide Foam Parts”, E. I. du Pont de Nemours and Company,
(2010).
130. ASTM Standard D3039, “Standard Test Method for Tensile Properties of Polymer Matrix
Composite Materials”, ASTM Internation, (2014).
131. A. Brooks, “NASA & Boeing Team up to Construct Composite Fuel Tank”, Composites
Today, http://www.compositestoday.com/2014/03/nasa-boeing-team-construct-
composite-fuel-tank/ , (2014).
132. J. Zhang, “Design of polymer composites with improved adhesion and wear properties”,
University of Illinois, 2008.
133. B. Piascik, “Materials, Structures, Mechanical Systems, and Manufacturing Roadmap,
Technology Area 12”, NASA, (2012).
134. J. Johnston, F. Li, F. Harris, T. Takekoshi, “Synthesis and characterization of imide
oligomers end-capped with 4-(phenylethynyl)phthalic anhydrides”, 35, 4865-4873,
(1994).
135. J. Moll, S. White, N. Sottos, “A self-sealing fiber-reinforced composite” J. of Compos.
Mater., 44, 2573-2585 (2010).
136. U. Gaur, B. Miller, “Effects of environmental exposure on fiber/epoxy interfacial shear
strength”, Polym. Compos., 11, 217-222, (2004).
137. B. Blaiszik, M. Baginska, S. White, N. Sottos, “Autonomic Recovery of Fiber/Matrix
Interfacial Bond Strength in a Model Composite” Adv. Funct. Mater., 20, 3547–355,
(2010).
138. R. Heydenreich, “Cryotanks in future vehicles”, Cryogenics, 38, 125-130, (1998).
139. B. Briscoe, S. Sinha, “Wear of polymers”, Proc. Inst. Mech. Eng., Part J: J. Eng. Tribol.
216, 401-413, (2002).
140. J. Hay, G. Pharr, “Instrumented Indentation Testing, ASM Handbook Volume 8:
Mechanical Testing and Evaluation, 10th ed.”, 231-243, (2000).
141. “Victrex PEEK 450G”, Victrex PLC, (2015)
142. “Dupont Vespel SP-1 Polyimide Isostatic Shapes”, E. I. du Pont de Nemours and
Company, (2014).
214
143. H. Brown, A. Yang, T. Russell, W. Volksen, E. Kramer, T. Russel, “Diffusion and Self-
Adhesion of the Polyimide PMDA-ODA” Polymer, 29, 1807 (1988).
144. A. Yamamoto, T. Hamada, I. Hisaki, M. Miyata, N. Tohnai, “Dynamically Deformable
Cube-like Hydrogen-Bonding Networks in Water-Responsive Diamondoid Porous
Organic Salts”, Angewandte Chemie International Edition, 52, 1709-1712, (2013).